This PDF file contains the front matter associated with SPIE Proceedings Volume 7442, including the Title Page, Copyright information, Table of Contents, and the Conference Committee listing.

Digital information and records are vital to the human race regardless of the nationalities and eras in which they were produced. Digital image contents are produced at a rapid pace from cultural heritages via digitalization, scientific and experimental data via high speed imaging sensors, national defense satellite images from governments, medical and healthcare imaging records from hospitals, personal collection of photos from digital cameras. With these mass amounts of precious and irreplaceable data and knowledge, what standards technologies can be applied to preserve and yet provide an interoperable framework for accessing the data across varieties of systems and devices? This paper presents an advanced digital image archival system by applying the international standard of MPEG technologies to preserve digital image content.

In this paper, we describe a novel technique for vision based UAV (unmanned aerial vehicle) navigation. In this technique, the navigation (position estimation) problem is formulated as a tracking problem and solved by a particle filter. The state and observation models of the particle filter are established based on a stereo analysis of the image sequence generated by the UAV's video camera in connection with a DEM (digital elevation map) of the area of the flight, which helps to control estimation error accumulation. The efficacy of this technique is demonstrated by simulation experimental results.

Most modern tracking techniques assume that the object comprises a large percentage of the image frame, however when the object is contained in a small number of pixels tracking via feature based methods is difficult, because they require a dense feature set which does not exist within small regions. As an alternative to dynamic boundary based methods, which require only a boundary between the object and the background, but often fail in busy enviroments, we propose using a novel graph cuts implemenation to obtain a more robust segmentation. The push-relabel method was chosen because of its lower time complexity. In addition the algorithm was expanded to the RGB color-space. This is done by a probabilistic combination of the RGB pixel values. This addition, by using all the information captured by the camera, allow objects with similar appearances and objects with large variances in color to be segmented. The final addition made to the the push-relabel algorithm is an min-cut approximation method which runs in O(n) time. We show that this formulation of the graph cut algorithm allows for a fast and accurate segmentation at 30 frames per second.

Alpha-rooted phase correlation (ARPC) is a recently-developed variant of classical phase correlation that includes a Fourier domain image enhancement operation. ARPC combines classical phase correlation with alpha-rooting to provide tunable image enhancement. The alpha-rooting parameters may be adjusted to provide a tradeoff between height and width of the ARPC main lobe. A high narrow main lobe peak provides high matching accuracy for aligned images, but reduced matching performance for misaligned logos. A lower, wider peak trades matching accuracy on aligned logos, for improved matching performance on misaligned imagery. Previously, we developed ARPC and used it in the spatial domain for logo recognition as part of an overall automated document analysis problem. However, spatial domain ARPC performance can be sensitive to logo misalignments, including rotational misalignment. In this paper we use ARPC as a match metric in the radon transform domain for logo recognition. In the radon transform domain, rotational misalignments correspond to translations in the radon transform angle parameter. These translations are captured by ARPC, thereby producing rotation-invariant logo matching. In the paper, we first present an overview of ARPC, and then describe the logo matching algorithm. We present numerical performance results demonstrating matching tolerance to rotational misalignments. We demonstrate robustness of the radon transform domain rotation estimation to noise. We present logo verification and recognition performance results using the proposed approach on a public domain logo database. We compare performance results to performance obtained using spatial domain ARPC, and state-of-the-art SURF features, for logos in salt-and-pepper noise.

Multibiometric systems offer more reliable and accurate performance combining the benefits of using multiple traits for user authentication. Due to incompatible biometric characteristics such as unmatched image patterns, improper feature registration and feature space representation, image scaling and unfeasible fusion schemes often degrades the performance of multibiometric systems. This paper focuses on the benefits of feature level and match score level fusions of face and ear biometrics using scale invariant feature transform (SIFT) representation and probabilistic graph. The proposed fusion techniques first compute and detect the SIFT features from face and ear images independently. Further probabilistic graphs are drawn on extracted feature points. By using iterative relaxation algorithm in both the graphs, which are drawn on face and ear images, corresponding feature points are searched and match points are paired and grouped into two independent sets. During feature level fusion, both the feature sets are concatenated together into an augmented group. Combined feature set is normalized using 'min-max' normalization rule and finally the concatenated feature vector is used for verification. In match score level fusion, independent verifications are performed using relaxation based probabilistic graphs and point pattern matching algorithm. As a result, independent matching scores generated from face and ear biometrics is fused together using 'sum' rule. The reported experimental results show the performance improvements in verification by applying feature level. and score level fusions.

In this paper, fusion of Principal Component Analysis (PCA) and generalization of Linear Discriminant Analysis (LDA) in the context of multiview face recognition is proposed. The generalization of LDA is extended to establish correlation between face classes in the transformed representation, which is called canonical covariate. The proposed work uses Gabor filter bank for extracting facial features characterized by spatial frequency, spatial locality and orientation to compensate the variations in face that occur due to change in illumination, pose and facial expression. Convolution of Gabor filter bank with face images produces Gabor face representations with high dimensional feature vectors. PCA and canonical covariate are then applied on the Gabor face representations to reduce the high dimensional feature spaces into low dimensional Gabor eigenfaces and Gabor canonical faces. Reduced eigenface vector and canonical face vector are fused together using weighted mean fusion rule. Finally, support vector machines have been trained with augmented fused set of features to perform recognition task. The proposed system has been evaluated with UMIST face database and performs with higher recognition accuracy for multi-view face images.

H.264/AVC encoder complexity is mainly due to variable block size in Intra and Inter frames. This makes H.264/AVC very difficult to implement, especially for real time applications and mobile devices. The current technological challenge is to conserve the compression capacity and quality that H.264 offers but reduce the encoding time and, therefore, the processing complexity. This paper applies machine learning technique for video encoding mode decisions and investigates ways to improve the process of generating more general low complexity H.264/AVC video encoders. The proposed H.264 encoding method decreases the complexity in the mode decision inside the Inter frames. Results show, on average, a 67.36% reduction in encoding time, a 0.2 dB decrease in PSNR, and an average bit rate increase of 0.05%.

Wireless video transmission is prone to potentially low data rates and unpredictable degradations due to timevarying channel conditions. Such degradations are difficult to overcome using conventional video coding techniques. Scalable video coding offers a flexible bitstream that can be dynamically adapted to fit the prevailing channel conditions. Advances in scalable video compression techniques, such as the newly adopted scalable extension of H.264/AVC, as well as recent advances in wireless access technologies offer possibilities for tackling this challenge. In this paper, a content-aware scheduling and resource allocation scheme is proposed, that uses a gradient-based scheduling framework in conjunction with scalable video coding techniques to provide multiple high quality video streams over a range of operating conditions to multiple users. Simulation results show that the proposed scheme performs better than conventional content-independent scheduling techniques.

Performance anomaly is a serious problem that would affect the throughput in multirate IEEE 802.11 wireless networks. Existing performance anomaly solutions improve the throughput by adjusting one or more MAC parameters. However, they do not consider the frame delay. Thus the frame delay of some stations would be increased due to the MAC parameter adjustments. The increase of frame delay would seriously affect the quality of service (QoS) of multimedia applications such as video communications. In this paper we classify the existing solutions based on the MAC parameters to adjust, and analyze the effect of each type of solution on the frame delay. Numerical results show that the effects on frame delay significantly differ between the various types of solutions. This provides a novel perspective to evaluate the existing performance anomaly solutions.

The problem of accurate motion vector (MV) prediction with a small search window size for memory-constrained systems is studied in this work. The spatial and temporal correlations of MVs are first investigated. Then, two effective search window prediction algorithms are developed to determine the location, size, shape of a search window for each macroblock. The trade-off between the search window size/shape and distortion is analyzed. Finally, the efficiency of the proposed algorithms is demonstrated by experimental results.

Bandwidth-efficient video file synchronization between remote clients is an important task. When heterogeneous mobile clients want to synchronize their local video data to that of a remote party at a desired resolution and distortion level, it is wasteful and unnecessary to retransmit the entire video data, especially when the differences are minor while the clients are limited in transmission bandwidth. We present VSYNC (video-sync), an incremental video file synchronization protocol that automatically detects and transmits differences between the video files without prior knowledge of what is different. VSYNC generalizes the popular universal file synchronization tool rsync to a semantics-aware utility that handles synchronization of video data. An important attribute of VSYNC is that it allows synchronization to within some quantitative distortion constraint. VSYNC can be easily embedded in a codec or transcoder, and can be used to synchronize videos encoded with different parameters or stored in different, possibly proprietary, formats. A hierarchical hashing scheme is designed to compare the video content at the remote ends, while a lossy distributed video coding framework is used to realize compression gains in the update steps. Experimental results of three heterogeneous mobile clients synchronizing to an updated video file at the remote server validate the performance gains in rate-savings attained by VSYNC compared to directly sending the updated video files using H.26x or synchronizing using universal file synchronization protocols such as rsync.

In this paper, we propose an automatic image browsing method based on image categorization to effectively browse high-resolution images on small-display-devices such as cellular phones and digital cameras. Based on face detection algorithm and spectrum analysis, input images are categorized into face images and non-face images. The non-face images are again categorized into close-up view images and non-close-up view images. For the non-close-up view images, we conduct further classification into images-with-vanishing-point and images-without-vanishing-point. In the face images case, the browsing path is determined by face locations. In images with vanishing point case, the path is decided along the vanishing lines, while in case of images without vanishing point, we detect the saliency regions using color variance and edges, and the browsing route is determined by the location of saliency regions. We estimate the accuracy of the proposed image classification algorithm through experiments. Subjective evaluation is also conducted to assess the proposed system for automatic image browsing. Experimental results indicate that our system increases viewing satisfaction to small-display viewers.

We describe a solution for image restoration in a computational camera known as an extended depth of field (EDOF) system. The specially-designed optics produce point spread functions that are roughly invariant with object distance in a range. However, this invariance involves a trade-off with the peak sharpness of the lens. The lens blur is a function of lens field-height, and the imaging sensor introduces signal-dependent noise. In this context, the principal contributions of this paper are: a) the modeling of the EDOF focus recovery problem; and b) the adaptive EDOF focus recovery approach, operating in signal-dependent noise. The focus recovery solution is adaptive to complexities of an EDOF imaging system, and performs a joint deblurring and noise suppression. It also adapts to imaging conditions by accounting for the state of the sensor (e.g., low-light conditions).

Today's video codecs have evolved primarily to meet the requirements of the motion picture and broadcast industries, where high‐complexity studio encoding can be utilized to create highly‐compressed master copies that are then broadcast one‐way for playback using less‐expensive, lower‐complexity consumer devices for decoding and playback. Related standards activities have largely ignored the computational complexity and bandwidth constraints of wireless or Internet based real‐time video communications using devices such as cell phones or webcams. Telecommunications industry efforts to develop and standardize video codecs for applications such as video telephony and video conferencing have not yielded image size, quality, and frame‐rate performance that match today's consumer expectations and market requirements for Internet and mobile video services. This paper reviews the constraints and the corresponding video codec requirements imposed by real-time, 2-way mobile video applications. Several promising elements of a new mobile video codec architecture are identified, and more comprehensive computational complexity metrics and video quality metrics are proposed in order to support the design, testing, and standardization of these new mobile video codecs.

The paper presents the results of a preliminary experiment in which we investigated some of the cognitive factors involved in viewing a video clip such as comprehension and cognitive load. The aim of our work is to have a measure which can help establish what visual quality is required by users to successfully assimilate and understand the informational content of a video clip with a minimum of user cost.

In this paper a procedure for subjective evaluation of the new JPEG XR codec for compression of still pictures is described in details. The new algorithm has been compared to the existing JPEG and JPEG 2000 standards when considering compression of high resolution 24 bpp pictures, by mean of a campaign of subjective quality assessment tests which followed the guidelines defined by the AIC JPEG ah-hoc group. Sixteen subjects took part in experiments at EPFL and each subject participated in four test sessions, scoring a total of 208 test stimuli. A detailed procedure for statistical analysis of subjective data is also proposed and performed. The obtained results show high consistency and allow an accurate comparison of codec performance.

JPEG XR, a new international standard for image coding, was approved as ITU-T Recommendation T.832 in March 2009, and as ISO/IEC international standard 29199-2 in July 2009. JPEG XR was designed based on Microsoft coding technology known as HD Photo. Since JPEG XR is an emerging new specification, exploration of advanced encoding techniques for JPEG XR is an important area of study. In order to advance understanding of JPEG XR and its capabilities, the development of enhanced encoding techniques for optimization of encoded JPEG XR perceptual image quality is particularly valuable. This paper presents techniques and results focusing on exploring the capabilities of the spatially adaptive quantization syntax of the emerging JPEG XR standard.

Traditionally, the quality of the human vision is evaluated by a subjective test in which the examiner asks the patient to read a series of characters of different sizes, located at a certain distance of the patient. Typically, we need to ensure a subtended angle of vision of 5 minutes, which implies an object of 8.8 mm high located at 6 meters (normal or 20/20 visual acuity). These characters constitute what is known as the Snellen chart, universally used to evaluate the spatial resolution of the human eyes. The mentioned process of identification of characters is carried out by means of the eye - brain system, giving an evaluation of the subjective visual performance. In this work we consider the eye as an isolated image-forming system, and show that it is possible to isolate the function of the eye from that of the brain in this process. By knowing the impulse response of the eye´s system we can obtain, in advance, the image of the Snellen chart simultaneously. From this information, we obtain the objective performance of the eye as the optical system under test. This type of results might help to detect anomalous situations of the human vision, like the so called "cerebral myopia".

Most current TV broadcasting systems use the interlaced scanning scheme in order to reduce bandwidth. However, this interlaced scan may introduce undesirable artifacts. A large number of deinterlacing methods have been proposed. Traditionally, the performance of deinterlacing methods has been measured in terms of pixel differences. However, there are some impairments which may not be accurately reflected in PSNR or MSE. In this paper, we choose several deinterlacing methods and apply them to a set of video sequences. Then, we perform subjective tests for the deinterlaced video sequences. The perceptual video quality scores are presented along with some discussions.

During a cognitive stimulation session where elders with cognitive decline perform stimulation activities, such as solving puzzles, we observed that they require constant supervision and support from their caregivers, and caregivers must be able to monitor the stimulation activity of more than one patient at a time. In this paper, aiming at providing support for the caregiver, we developed a vision-based system using an Phase-SDF filter that generates a composite reference image which is correlated to a captured wooden-puzzle image. The output correlation value allows to automatically verify the progress on the puzzle solving task, and to assess its completeness and correctness.

The extension of video applications to enable 3D perception, which typically is considered to include a stereo viewing experience, is emerging as a mass market phenomenon, as is evident from the recent prevalence of 3D major cinema title releases. For high quality 3D video to become a commonplace user experience beyond limited cinema distribution, adoption of an interoperable coded 3D digital video format will be needed. Stereo-view video can also be studied as a special case of the more general technologies of multiview and "free-viewpoint" video systems. The history of standardization work on this topic is actually richer than people may typically realize. The ISO/IEC Moving Picture Experts Group (MPEG), in particular, has been developing interoperability standards to specify various such coding schemes since the advent of digital video as we know it. More recently, the ITU-T Visual Coding Experts Group (VCEG) has been involved as well in the Joint Video Team (JVT) work on development of 3D features for H.264/14496-10 Advanced Video Coding, including Multiview Video Coding (MVC) extensions. This paper surveys the prior, ongoing, and anticipated future standardization efforts on this subject to provide an overview and historical perspective on feasible approaches to 3D and multiview video coding.

To facilitate new video applications such as three-dimensional video (3DV) and free-viewpoint video (FVV), multiple view plus depth format (MVD), which consists of both video views and the corresponding per-pixel depth images, is being investigated. Virtual views can be generated using depth image based rendering (DIBR), which takes video and the corresponding depth images as input. This paper discusses view synthesis techniques based on DIBR, which includes forward warping, blending and hole filling. Especially, we will emphasize on the techniques brought to the MPEG view synthesis reference software (VSRS). Unlike the case in the field of computer graphics, the ground truth depth images for nature content are very difficult to obtain. The estimated depth images used for view synthesis typically contain different types of noises. Some robust synthesis modes to combat against the depth errors are also presented in this paper. In addition, we briefly discuss how to use synthesis techniques with minor modifications to generate the occlusion layer information for layered depth video (LDV) data, which is another potential format for 3DV applications.

There is significant industry activity on delivery of 3D video to the home. It is expected that 3D capable devices will be able to provide consumers with the ability to adjust the depth perceived for stereo content. This paper provides an overview of related techniques and evaluates the effectiveness of several approaches. Practical considerations are also discussed.

In surveillance, reconnaissance and numerous other video applications, enhancing the resolution and quality enhances the usability of captured video. In many such applications, video is often acquired from low cost legacy sensors that offer low resolution due to modest optics and low-resolution arrays, providing imagery that may be grainy and missing important details - and still face transmission bottlenecks. Many post-processing techniques have been proposed to enhance the quality of the video and superresolution is one such technique. In this paper, we extend previous work on a real-time superresolution application implemented in ASIC/FPGA hardware. A gradient based technique is used to register the frames at the sub-pixel level. Once we get the high resolution grid, we use an improved regularization technique in which the image is iteratively modified by applying back-projection to get a sharp and undistorted image. The matlab/simulink proven algorithm was migrated to hardware, to achieve 320x240 -> 1280x960, at more than 38 fps, a stunning superresolution by 16X in total pixels. This significant advance beyond real-time is the main contribution of this paper. Additionally the algorithm is implemented in C to achieve real-time performance in software with optimization for Intel I7 processor. Fixed 32 bit processing structure is used to achieve easy migration across platforms. This gives us a fine balance between the quality and performance. The proposed system is robust and highly efficient. Superresolution greatly decreases camera jitter to deliver a smooth, stabilized, high quality video.

To achieve the high coding efficiency the H.264/AVC standard offers, the encoding process quickly becomes computationally demanding. One of the most intensive encoding phases is motion estimation. Even modern CPUs struggle to process high-definition video sequences in real-time. While personal computers are typically equipped with powerful Graphics Processing Units (GPUs) to accelerate graphics operations, these GPUs lie dormant when encoding a video sequence. Furthermore, recent developments show more and more computer configurations come with multiple GPUs. However, no existing GPU-enabled motion estimation architectures target multiple GPUs. In addition, these architectures provide no early-out behavior nor can they enforce a specific processing order. We developed a motion search architecture, capable of executing motion estimation and partitioning for an H.264/AVC sequence entirely on the GPU using the NVIDIA CUDA (Compute Unified Device Architecture) platform. This paper describes our architecture and presents a novel job scheduling system we designed, making it possible to control the GPU in a flexible way. This job scheduling system can enforce real-time demands of the video encoder by prioritizing calculations and providing an early-out mode. Furthermore, the job scheduling system allows the use of multiple GPUs in one computer system and efficient load balancing of the motion search over these GPUs. This paper focuses on the execution speed of the novel job scheduling system on both single and multi-GPU systems. Initial results show that real-time full motion search of 720p high-definition content is possible with a 32 by 32 search window running on a system with four GPUs.

Deinterlacing is a process of converting interlaced video signals into progressive video signals. Traditionally, deinterlacing methods have been evaluated in terms of PSNR. On the other hand, some artifacts can severely degrade perceptual video quality, but may not produce large mean square errors. One of such artifacts is flickering. This flickering artifact occurs more frequently as the video size increases. For example, for HD TV signals, this flickering artifact is more frequently observed. In this paper, we propose a deinterlacing method which noticeably reduces this flickering artifact with slight loss in terms of PSNR.

Due to its coding efficiency and strong demand for compression technology from the professional video production industry, the JVT recently standardized five new profiles for high quality video - one predictive and the rest intra only. Of these, the new High 4:4:4 Profile sits atop virtually the entire original chain of H.264/AVC profiles (not counting the more recent scalable and multiview coding profiles, which are somewhat specialized), and subsumes the others -- it lacks only the error resilience tools found in Baseline and Extended Profiles. It may thus be regarded as the ultimate codec in the new standard family. While several companies are known to have developed internal technology for this profile, to our knowledge no commercial product has been released to support this profile. In this paper, we present some elements of the first commercial software encoder for the H.264/AVC High 4:4:4 Profile. We make significant progress on realtime performance, using the new Intel i7 processor, while providing some visual performance results on a potential application to infrared sensors, where high dynamic range is critical.

This paper offers an introduction to a new paradigm for resampling filters (both downsampling and upsampling filters), with an emphasis on a new performance metric - the roundtrip (down, up) high resolution reconstruction quality. Existing philosophy in the construction of resampling filters is to design them individually to minimize aliasing artifacts caused by the resampling process. The net result of a roundtrip high-to-low-to-high resolution reconstruction while using such filters certainly minimizes aliasing, but misses an important opportunity -- to capture more information in the lower resolution signal which, though aliased, can contribute to higher quality reconstruction when paired with an appropriately designed upsampling filter. While the criterion of high-resolution reconstruction quality is not new, alias control has been so heavily emphasized to date that surprisingly little work has been done in the literature to specifically address this important metric. We provide a setting for this new theory in the context of Laplace Pyramids, and develop specific resampling filters that outperform previous state-of-the-art filters in our context. Our filters were first developed and proposed in the course of the ISO/ITU Joint Video Team's Scalable Video Coding (SVC) standardization project.

We review design of 4-, 8-, and 16-point transforms currently used in image and video coding standards, and compare them with fast implementations of Discrete Cosine Transform of various other sizes (including non-dyadic even and odd numbers) in the range of 2-64. We show that among such transforms there exist few that offer better complexity/coding gain tradeoffs than current dyadic-sized transforms. In our construction and analysis we utilize an array of known techniques (such as Heideman's mapping between DCT and DFT, Winograd short length DFT modules, prime-factorand common-factor algorithms), and also offer a new factorization scheme for even-sized scaled transforms.

A computationally efficient adaptive 2-stage Karhunen-Loeve Transform (KLT) scheme for spectral decorrelation in hyperspectal lossy bandwidth compression is presented. The component decorrelation of the JPEG 2000 (extension 2) is replaced with the proposed adaptive 2-stage KLT spectral decorrelation scheme. Direct application of a single KLT across the entire set of hyperspectal imagery may not be computationally practical. The proposed scheme alleviates this problem by partitioning the spectral data set into small subsets. The spectral correlation within each partition is removed via the 1st-stage KLT operation. To remove the remaining inter-partition correlation, a 2nd-stage KLT is applied to the top few sets of eaui-level principal component (PC) images from the 1st-stage. The computation savings resulting from 2-stage KLT is parametrically quantified. The proposed adaptive 2-stage KLT uses only a fraction of the equi-level 1st-stage PC images in the 2nd-stage KLT process. This adaptive scheme results in reducing the size of the 2nd-stage KLT transformation matrices and further improvement in computational complexity and coding efficiency. It is shown that reconstructed image quality, as measured via statistical and/or machine-based exploitation measures, is improved by using a smaller partition size in the 1st-stage KLT. A criterion based on the components of the eigenvectors of the cross-covariance matrix is established to identify such 1st-stage PC images. The proposed adaptive spectral decorrelation scheme also reduces the overhead bits required to transmit the covariance matrices, or eigenvectors, along the coding bit stream to the receiver through the downlink channel.

In this paper, a rate-distortion optimized quantization scheme is described with application to H.264 video encoding. An efficient implementation of H.264 macroblock level adaptive quantization parameter selection is also described. Together these two encoder-only changes can achieve on average over 6% bit rate reduction under common testing conditions that are used in the H.264 standardization community. The described techniques provide this improvement in compression capability while retaining conformance of the encoded data to the H.264 standard. Thus, full compatibility with standard decoders can be achieved when applying these techniques.

JPEG2000, the international standard for still image compression, uses wavelet transform. JPEG2000 uses the 9/7 Daubechies filter and the 5/3 Le Gall filter for transformation. In this paper, we present a method for image mirroring and rotation in the wavelet domain, by manipulating the transformation coefficients. Our technique requires no additional complexity and the perfect reconstruction property is preserved. The paper also discusses a method to achieve image interpolation, which constructs a higher resolution image from a lower resolution one. The wavelet domain coefficients of the image are utilized for the same. For comparison purposes, the low frequency components obtained after the wavelet transform of the image are considered as the low resolution image. Then we predict the rest of the components in order to reconstruct the higher resolution image. The reconstructed image is then compared with the original image.

The ISO standard JPEG 2000 Part 9 (15444-9) specifies a versatile and flexible image browsing and delivering protocol that allows the interactive selection of regions of large images and their transmission over a narrow bandwidth connection. However, due to the enormous flexibility, achieving interoperability between software from differing vendors is not an easy task. To address this challenge, the JPEG committee started an initiative in the form of an amendment to 15444-9 to establish common grounds on which interoperability can be defined. The outcome of this work are recommendations which subsets of JPIP vendors should focus on, hopefully easing the adoption of JPIP by identifying the options the committee found in widespread use. In this paper, the design and evolution of JPIP interoperability will be discussed, the grounds on which interoperability can be achieved- variants and profiles- will be introduced, and their design will be motivated. The paper closes with an outlook how to extend this amendment for future applications.

Here we will present challenges to be met in connection with the application of spectroscopic imaging and in particular Raman microspectroscopy for life sciences and biomedicine. We start with an introduction of a combinatorial approach between florescence imaging, Raman microspectroscopy and innovative statistical Raman data analysis methods for rapid diagnosis and prevention of infectious diseases. Furthermore we will report about the multimodal application of Raman and CARS imaging for an early diagnosis of cancer.

Mathematical morphology is a powerful tool for filtering, segmentation and texture analysis, extended to multivariate signal in the last years. The major limitations of applying it to colour image reside in the non-linear behavior of the Human Visual System to the perception of colour. So a direct extension of the multivariate approach to colour image is not appropriate and the existing approaches can not offer generic solutions from a perceptual point of view. To overcome this limit, we present a coherent solution for the addition/subtraction parts of the colour dilatation/erosion specification, which didn't limit the structural element to the flat ones. By combination of two perceptual colour spaces, we define a partial order, specified by a perceptual colour distance. By this way, we solve lot of problems induced by all methods based on bit mixing or lexicographical strategies. In addition, we define unic supremum and infimum in the colour space allowing classical developments for filtering or segmentation and colour texture analysis without colour artefacts. In this paper, we will discuss about the colour spaces and their specificities, then we present the possible colour ordering schemes for mathematical morphology. In a second time, we develop our specific approach, beginning by the discussion about an adapted colour space, following with the extrema extraction formulation in this adapted colour space, by distance computation. Then we propose the colour addition expression needed in the complete formulation of supremum and infimum. Finally, we show the first results in colour textured image filtering and we conclude with perspectives.

This paper develops a digital decoding design for the imaging system with phase coded lens. The phase coded lens is employed to extend the depth of filed (DoF), and the proposed design is used to restore the special-purpose blur caused by the lens. Since in practice the imaging system inevitably contains manufacturing inaccuracy, it is often difficult to obtain precise point spread function (PSF) for image restoration. To deal with this problem, we develop a flow for designing filters without PSF information. The imaging system first takes a shot of a well-designed test chart to have a blur image of the chart. This blur image is then corrected by using the perspective transformation. We use both of the image of the test chart and the corrected blur image to calculate a minimum mean square error (MMSE) filter, so that the blur image processed by the filter can be very alike to the test chart image. The filter is applied to other images captured by the imaging system in order to verify its effectiveness in reducing the blur and for showing the capability of extending the DoF of the integrated system.

In this paper, a method integrating image processing and signal frequency is used to estimate relative misalignment of micro-structure the Double-sided Lenticular Lens Films (LLF). The pitch of the convex on each side of the LLF is 900μm. The main steps of this research consist of image denoising, edge detection and establishment of the misalignment v.s. image magnitude relation. A Butterworth low-pass filter is used to eliminate the noise due to debrits on the polymer substrate. The images that are acquired from LLFs for different relative misalignment quantities are used to analyze the spectrum and get its relation between misalignment position and frequency energy distribution. The experiment shows the proposed method may distinguish between cases with different relative misalignment of microstructure on LLF with a resolution of 20μm.

In this paper we present a new method for superresolution of depth video sequences using high resolution color video. Here we assume that the depth sequence does not contain outlier points which can be present in the depth images. Our method is based on multiresolution decomposition, and uses multiple frames to search for a most similar depth segments to improve the resolution of the current frame. First step is the wavelet decomposition of both color and depth images. Scaling images of the depth wavelet decomposition, are superresolved using previous and future frames of the depth video sequence, due to their different nature. On the other side wavelet band are improved using both previous frames of the wavelet bands and wavelet bands of color images since similar edges might appear in both images. Our method shows significant improvements over some recent depth images interpolation methods.

The cavernous nerves course along the surface of the prostate and are responsible for erectile function. Improvements in identification, imaging, and visualization of the cavernous nerves during prostate cancer surgery may improve nerve preservation and postoperative sexual potency. In this study, 2-D OCT images of the rat prostate were segmented to differentiate the cavernous nerves from the prostate gland. Three image features were employed: Gabor filter, Daubechies wavelet, and Laws filter. The features were segmented using a nearestneighbor classifier. N-ary morphological post-processing was used to remove small voids. The cavernous nerves were differentiated from the prostate gland with a segmentation error rate of only 0.058 ± 0.019.

It is well known that the Zernike expansion of the wavefront aberrations has been extensively used to evaluate the performance of image forming optical systems. Recently, these techniques were adopted in the field of Ophthalmology to evaluate the objective performance of the human ocular system. We have been working in the characterization and evaluation of the performance of normal human eyes; i.e., eyes which do not require any refractive correction (20/20 visual acuity). These data provide us a reference model to analyze Pre- and Post-Operated results from eyes that have been subjected to laser refractive surgery. Two different ablation techniques are analyzed in this work. These techniques were designed to correct the typical refractive errors known as myopia, hyperopia, and presbyopia. When applied to the corneal surface, these techniques provide a focal shift and, in principle, an improvement of the visual performance. These features can be suitably described in terms of the PSF and MTF of the corresponding Pre- and Post-Operated wavefront aberrations. We show the preliminary results of our comparison.

This paper demonstrates a technique that could prove useful for extracting three-dimensional (3D) models from a single two-dimensional (2D) digital in-line holographic microscopy (DIHM) recording. Multiple intensity images are reconstructed at a range of depths through a transmissive or partially transmissive scene recorded by DIHM. A two step segmentation of each of these reconstructed intensity images facilitates the construction of a data set of surfaces in 3D. First an adaptive thresholding step and then a border following step are implemented. The surfaces of segmented features are rendered in 3D by applying the marching cubes algorithm to polygonize the data set. Experimental results for a real world DIHM capture of a transmissive glass sample are presented to demonstrate this segmentation and visualization process.

Classical correlation filters for object detection and location estimation are designed assuming that the appearance and the shape of the target are explicitly known. In this work we assume that the target is given at unknown coordinates in a reference image corrupted by additive noise. Optimal correlation filters, with respect to signal-to-noise ratio and peak-to-output energy, for object detection and location estimation are derived. Two mathematical models of observed images are used; the additive noise model for the reference image and the non-overlapping background model for the input scene. Computer simulation results obtained with the proposed filters are presented and compared with those of common correlation filters.

Color temperature(CT) conversion of triprimary color display from one white point to another on the Planckian's locus with the maximal brightness has been proposed. However, whether converting an original white point to another white point on the isotemperature line will enlarge maximal brightness more than converting an original white point to another white point on the Planckian's locus needs to be determined. This paper proposes a new algorithm to enlarge maximal brightness by calling the center of gravity method of color mixing in the acceptable color difference range while the CT is converted. From the prior study, we find that the apexes of color gamut boundary move along the line of center gravity of primaries while the total brightness varies, where the line of center gravity of primaries is linked by the color points mixed by two or more full primary colors and one partial primary color. And in CIE 1931 color space, the color gamut boundary expanding from white point as total brightness decreasing will touch the isotemperature line with its apexes. Therefore, the best point of CT conversion of tri-primary color display with more maximal brightness is determined by the isotemperature line and the line of center gravity of primaries. Further, the theory extends to multi-primary color displays. Lastly, the simulations prove that converting a white point to another on the isotemperature line enlarges maximal brightness more than converting a white point to another on Planck's locus.

This paper proposes the use of the Hilbert transform for analysis of fringe patterns, which are obtained in Michelson Interferometer. The fringe patterns are evaluated by Continuous wavelet transform and Hilbert transform algorithms to obtain phase map and refractive index. Continuous wavelet transform algorithm shows a better filtering effect. However Hilbert transform algorithm gives a noisy result, since its frequency spectrum has harmonic noise. Results show that the phase map is affected by Hilbert Filter.

We have been working in the interferometric analysis of the ablation profile obtained with different techniques of refractive surgery, applied directly on hard contact lenses. We have demonstrated qualitatively that different ablations produce different fringe patterns; implying different focal shifts1. These results were obtained by means of a Mach- Zehnder type interferometer, where we used a similar unablated contact lens as a reference. Due to the size of each sample, it is difficult to get different fringe patterns with different phase factors. Therefore, the typical phase shifting methods are not suitable in our case. To determine the corresponding profile caused by the different ablation techniques we applied in this work the interpolation method that provide an analysis of static fringe patterns. This method of phase retrieval allows us to obtain the PSF and MTF related to each profile. The advantage of this procedure is that we can obtain a time invariant performance of the resulting ablated surface.

A digital image processing approach to develop an automatic method for the objective measure of some thread parameters in a non closely-woven fabric is presented. The parameters addressed by this method are thickness, periodicity, regularity, area of holes between threads and cover factor. Applied techniques range from image thresholding for the segmentation of the threads from the background, to morphological operations such as skeletonization and non-lineal filters. They are successfully applied to segment vertical and horizontal lines describing the geometry of the threads. The mentioned parameters are derived from further labeling of the skeletonized images.

The image block matching algorithm based on motion vectors of correlative pixels in oblique direction is presented for digital image stabilization. The digital image stabilization is a new generation of image stabilization technique which can obtains the information of relative motion among frames of dynamic image sequences by the method of digital image processing. In this method the matching parameters are calculated from the vectors projected in the oblique direction. The matching parameters based on the vectors contain the information of vectors in transverse and vertical direction in the image blocks at the same time. So the better matching information can be obtained after making correlative operation in the oblique direction. And an iterative weighted least square method is used to eliminate the error of block matching. The weights are related with the pixels' rotational angle. The center of rotation and the global emotion estimation of the shaking image can be obtained by the weighted least square from the estimation of each block chosen evenly from the image. Then, the shaking image can be stabilized with the center of rotation and the global emotion estimation. Also, the algorithm can run at real time by the method of simulated annealing in searching method of block matching. An image processing system based on DSP was used to exam this algorithm. The core processor in the DSP system is TMS320C6416 of TI, and the CCD camera with definition of 720×576 pixels was chosen as the input video signal. Experimental results show that the algorithm can be performed at the real time processing system and have an accurate matching precision.

A multi-function lens test instrument is report in this paper. This system can evaluate the image resolution, image quality, depth of field, image distortion and light intensity distribution of the tested lens by changing the tested patterns. This system consists of a tested lens, a CCD camera, a linear motorized stage, a system fixture, an observer LCD monitor, and a notebook for pattern providing. The LCD monitor displays a serious of specified tested patterns sent by the notebook. Then each displayed pattern goes through the tested lens and images in the CCD camera sensor. Consequently, the system can evaluate the performance of the tested lens by analyzing the image of CCD camera with special designed software. The major advantage of this system is that it can complete whole test quickly without interruption due to part replacement, because the tested patterns are statically displayed on monitor and controlled by the notebook.

Recently, a blind image restoration algorithm based on camera microscanning was proposed. Unfortunately, the computational complexity of the algorithm is very high. In this paper we propose a fast algorithm for image restoration using the information obtained during camera microscanning. First, the captured observed images are decomposed into a pyramidal set of small images. Next, the blind iterative algorithm is applied for restoration of the set of small images. Finally, the resultant output image is constructed from the set of restored small images. Simulation results are presented and discussed. Preliminary results show that the processing time is significantly reduced.

Expensive measuring equipments and to some extent also professional digital still cameras are capable to capture scenes with high accuracy in representation of colors and consequently luminance. On the contrary, low-end equipments are in most of the cases adjusted by the manufacturer to produce visually attractive images with weak interest to reach high fidelity, measurement device close, reproduction of colors. In this paper we present a simple way how to use a consumer level digital camera as a 2D luminance meter or imaging photometer. Our approach is based mostly on measurement of spectral response curves of photosensitive elements and evaluation of output response from the raw image data captured by the camera. Utilization of these characteristics leads to a better calibration of the camera as a luminance meter. The performance improvement while using the image calibration data of the image sensor and objective lens is verified. The potential utilization of a consumer level digital camera as a substitute for expensive measuring devices is discussed.

This paper deals with modeling of scientific and multimedia images in the wavelet domain. Images transformed into wavelet domain have a special shape of probability density function (PDF). Thus wavelet coefficients PDFs are usually modeled using generalized Laplacian PDF model (GLM), which is characterized by two parameters. The wavelet coefficients modeling can be more efficient, while the Gaussian mixture model (GMM) is utilized. GMM model is given by addition of at least two Gaussian PDFs with different standard deviations. There will be presented equation system derived by moment method for GMM models parameters estimation. The equation system was derived for an addition of two GMM models. So it is suitable for advanced denoising systems, where an addition of two GMM random variables is considered (e.g. dark current in astronomical images).

An enhanced wavelet-based compression scheme for satellite image is proposed in this paper. The Consultative Committee for Space Data System (CCSDS) presented a recommendation which utilizes the wavelet transform and the bit plane coder for satellite image compression. The bit plane coder used in the CCSDS recommendation encodes the coefficient block of bit planes one by one and then truncates the unnecessary bit plane coefficient blocks. By this way, the contexts of bit planes are not considered as the redundancy embedded data which may be compressed further. The proposed scheme uses a bit plane extractor to parse the differences of the original image data and its wavelet transformed coefficients. The output of bit plane extractor will be encoded by a run-length coder and will be sent to the communication channel with the CCSDS compressed data. Comparing with the recommendation of CCSDS, under a reasonable complexity, the subjective quality of the image will maintained or even better. In addition, the bit-rate can be further decreased from 85% to 95% of the CCSDS image compression recommendation at the similar objective quality level. By using the lower bit rate lossy mode compression and bit plane compensation, it is possible to obtain lower bit rate and higher quality image than which the higher bit rate lossy mode compression can achieve.

Surface defect detection is an important task of industrial inspection that has traditionally relied on trained human vision. Automated and objective inspection methods based on image analysis have played a decisive role in the industrial progress of the last decades. We propose a new unsupervised novelty detection method for defect segmentation in textures. It uses a multiresolution Gabor filter scheme and shows the following properties: no need of any defect-free references or a training stage; any adjustable parameters, and applicability to both random and periodic textures. We apply the odd part of Gabor filters to the sample image, analyze the details obtained at different scales and orientations, and extract a number of background texture features from the sample under inspection. In the analysis, we assume that the wavelet coefficients of pixels can be suitably fitted by Gaussian mixtures, more specifically, by combining two normal distributions. One of them would correspond to the background texture whereas the other would account for the defective area. Since all the information is obtained from the sample image itself, the threshold selection is robust against possible sample to sample fluctuations such as heterogeneities in the material, inplane positioning errors, scale variations and lack of homogeneous illumination. The efficacy of the statistical analysis is demonstrated. The method is applied to a variety of samples that exhibit either periodic or random texture. A comparison with other unsupervised method designed for defect segmentation in periodic textures is done.

The increasing prevalence of obesity suggests a need to develop a convenient, reliable and economical tool for assessment of this condition. Three-dimensional (3D) body surface imaging has emerged as an exciting technology for estimation of body composition. This paper presents a new 3D body imaging system, which was designed for enhanced portability, affordability, and functionality. In this system, stereo vision technology was used to satisfy the requirements for a simple hardware setup and fast image acquisitions. The portability of the system was created via a two-stand configuration, and the accuracy of body volume measurements was improved by customizing stereo matching and surface reconstruction algorithms that target specific problems in 3D body imaging. Body measurement functions dedicated to body composition assessment also were developed. The overall performance of the system was evaluated in human subjects by comparison to other conventional anthropometric methods, as well as air displacement plethysmography, for body fat assessment.

Moving object detection from videos is one of important problems in intelligent video surveillance. Our goal is to extract moving foreground objects from the background of a scene captured from a stationary camera. We propose methods to improve a background subtraction (BGS) algorithm effectively using the spatial coherence of foreground regions - spatial filtering and region-based BGS. The proposed spatial filtering is based on the spatial coherence around the pixel neighborhood. In the region-based BGS, pixel deviation values (from the background model) are aggregated in a region for region-wise detection. We demonstrate the improved performance using the proposed methods on several indoor/outdoor video sequences.

We propose an adaptive optics system for a lightweight remote sensing sensor. The phase diversity (PD) technique, in which known wavefronts (Phase Diversity) are applied to the optics and the inherent aberrations are estimated using the acquired images without a priori information, is a key to realizing the system. For the reduction of computing cost and the enhancement of the estimation accuracy of aberration, a spatial light modulator (SLM) is adopted not only for wavefront compensator but also for PD generator. The SLM produces arbitrary "aberration modes" that are each represented by a Zernike polynomial. Therefore, optimal phase diversities are applied to the optical system and particular modes are effectively obtained, which makes it possible to overcome the conventional PD generated by defocusing that describes only quadratic form and lacks information of a particular mode. In order to solve the complex inverse problem of phase diversity with low computing cost, a general regression neural network (GRNN) is used. Moreover, principal component analysis compresses the input data for GRNN by extracting information from collected images in Fourier space, and reduces computation cost considerably. The performance is validated by numerical simulation, and the result of experiment using SLM is described.

The principal aim of this paper is to present a general view of the special optical systems used for acquiring of the astronomical image data, which are commonly called WFC or UWFC (Ultra Wide Field Camera), and of their transfer characteristics. UWFC image data analysis is very difficult. Images obtained from UWFC systems are distorted by many various optical aberrations. The influence of optical aberrations increases towards margins of the field of view. These aberrations distort the point spread function of the optical system and rapidly cut the accuracy of measurements. UWFC systems frequently have so called spatially variant properties. This paper deals with simulation and modeling of the UWFC optical systems used in astronomy and their transfer characteristics.

The technique to estimate the depth and 3D shape of an object from the images of the same sample obtained at different focus settings is called shape from focus (SFF). Conventional SFF methods sum up the focus values within a small window of each pixel in the image. It produces a surface distortion effect, and an inaccurate depth map is obtained. In this paper, a fast and accurate SFF method based on averaging filter is proposed. We suggest that instead of averaging focus values, averaging depth values produces more accurate depth map. The experimental results demonstrate the effectiveness and the efficiency of the proposed method in comparison to the conventional methods.

Pulmonary nodule detection is a binary classification problem. The main objective is to classify nodule from the lung computed tomography (CT) images. The intra class variability is mainly due to the grey-level variance, texture differences and shape. The purpose of this study is to develop a novel nodule detection method which is based on Two-dimensional Principal Component Analysis (2DPCA). We extract the futures using 2DPCA from nodule candidate images. Nodule candidates are classified using threshold. The proposed method reduces False Positive (FP) rate. We tested the proposed algorithm by using Lung Imaging Database Consortium (LIDC) database of National Cancer Institute (NCI). The experimental results demonstrate the effectiveness and efficiency of the proposed method. The proposed method achieved 85.11% detection rate with 1.13 FPs per scan.

A cell is the structural and functional unit of all known living organisms, and its three-dimensional shape is an interesting research topic and having many applications in biology. Usually, cells are kept surrounded with some liquid materials on glass plates. In obtained image sequence, liquid material causes unwanted background in the images, and some virtual images due to the glass plates occurs, which makes difficulty to recover the three-dimensional shape of the cell. Therefore, conventional optical passive methods for three-dimensional shape recovery do not compute depth map accurately. The purpose of this work is to reconstruct three-dimensional shape of HeLa cell by applying shape from focus (SFF) method. SFF method is one of the optical passive methods to estimate three-dimensional shape by using focal information from image sequence. To overcome problems from transparency and reflection, transparent part is segmented from images by using the fact that background of the cell does not have focal point, and an original image sequence is divided into two image sequences for real and virtual part by finding two focused points in itself. For more accurate segmentation of the background part, the labeling method is used, and for automatically dividing an original image sequence into two image sequences, the iterative threshold selection method is used. The proposed approach is tested by using HeLa cell which is one of the most famous cells in biological research area. The experimental result demonstrates the effectiveness.

This paper presents methods for motion detection and estimation of objects in a scene captured by a moving camera. A new iterative algorithm is developed and presented for the correction of geometrical distortion caused by global motion in a scene. A binary hypotheses test is subsequently established to classify the pixels in the corrected image as either locally moving (object motion) or not moving (stationary). The developed method incorporates estimates of additive white Gaussian noise in all steps and is therefore more robust than simple change detection method.

In this paper, a novel approach for extraction of object perspectivity using a new elemental image array is proposed. Generally, if we want to change the perspectivity of objects, we should change the position of object or camera, forward or backward. But, it makes the information of object image be limited. In other words, as the angle of camera is changed, the viewing angle is also done. Here, it makes the information of object be limited. As the object size is the bigger than before. In contrast of it, background image or other object image of rate is the smaller. Not only the information of object is limited, but also it is difficult to array experiment sets. Accordingly, in this paper, the perspectivity of objects is realized by using a new elemental image array which is made change the pinhole points horizontally. If the pinhole point is changed, the matching pixel is also changed. As if picked up image distance is the closer than before. So the perspectivity is changed. That is, It makes the similarly effect of changing the distance vertically. And it makes other perspectivity which is different from conventional method. Of course, the information is a little changed, but the image information is not much different. And it is not limited by the angle of camera or viewing angle. Besides experiment sets are maintained. Show the feasibility of the proposed method, some experiments with test objects are carried out and the results are presented.

We present in this work a multifocus image fusion based on the Daubechies wavelet transform applied to multi-focus microscopy color images acquired by the bright-field reflection microscopy technique. The fusion scheme is based on the Daub(2) and Daub(4) kernels of the Daubechies family. The fusion scheme is implemented in each RGB channel. Experimental results are presented using metallic samples.

Many countries have developed different mobile TV standards such as DVB-H, DMB, MediaFLO, OneSeg, and ATSCM/ H. In this paper, we compare different mobile TV standards from the viewpoint of video and audio encoding, data encapsulation, transmission network, and advantages and disadvantages of each technology. Also we propose future mobile TV applications 3-D mobile TV to provide backward compatibility with DMB and better bit rates.

Images associated with underwater imaging systems are normally degraded by the intervening water medium. The imaging instrument records not only the signal of interest, i.e., the radiance diffusely reflected from underwater target, but also the radiance scattered into the field of view by water molecule and particulates. In order to improve the system performance, range gated underwater imaging system is used to enhance image quality and visibility in turbid conditions. Range gated imaging utilizes time discrimination to improve signal-to-backscattering noise ratio by rejecting backscattered light from the medium. The range gated underwater imaging system basically consists of a pulsed laser system, a control and synchronous logics and a high-speed gated camera. Because a laser is a highly coherent light source, speckle noise results from the randomly constructive or destructive interference of the scattered light rays will appear in the images obtained from the range gated underwater imaging system. The random granular speckle noise brings great difficulty for the image processing. So the formation causes of speckle noise are discussed and several different material objects under standard light source and laser are chosen to carry out speckle noise comparative analysis. And a multidirectional morphological filtering algorithm for reducing speckle noise is proposed by using the characteristics of morphology's multi-resolution analysis and fast-computing. In order to evaluate the method objectively, equivalent number and speckle index are introduced. The experimental results demonstrate that the approach that is adopted not only can reduce the speckle noise of the image effectively but also can preserve the feature detail efficiently.

Range gated underwater laser imaging technique can eliminate backscattering noise effectively. While the images associated with underwater imaging systems are normally degraded seriously by the intervening water medium. And the speckle noise is especially severe for the reason that we adopt the system based on intensified gate imaging technology. Well known causes of image degradation underwater include turbidity, particulate matters in the water column, and the interaction between light and medium as light travels through water. Consequently, using full image formation models to design restoration algorithms is more complex in water than in air because it's hard to get the values of the model parameters relating to water properties, e.g., attenuation and scattering coefficients. To improve the quality of the low signal-to-noise ratio images obtained through range gated laser imaging system, an enhancement algorithm is proposed. The main purpose of the algorithm proposed for processing underwater images is to filter out unwanted noises and remain desired signals. This algorithm is based on the principle of the least square error method, which fits discrete image data to continuous piecewise curves. To simply the fitting of image data, the interval of each row and column is subdivided into several subintervals. Then a curve is used to fit the image data within the subinterval. To merge two adjacent lines together, a weighting technique with a linear weighting factor is imposed. A series of experiments are carried out to study the effects of the algorithm. And the signal-to-noise ratio shows that the proposed algorithm can achieve high quality enhancement images.

Shape from focus (SFF) is a passive optical method for 3D shape recovery, which has numerous applications in machine vision, range segmentation, and video microscopy. This paper introduces a new algorithm for shape from focus (SFF) based on multidimensional scaling (MDS) analysis. In contrast to the conventional focus measures operators, a three dimensional neighborhood, enabling to capture the effect of pixels from previous as well as next frames on focus value, is considered for each pixel in the image volume. A similarity matrix is computed using Euclidean metric for the sequence of these 3D neighborhoods corresponding to each object point. This matrix is then provided as input to MDS algorithm. The monotonic regression is applied which computes the fitness of the approximated configuration by using stress function as the criterion for the fitness. The energy of the components in lower dimensions is employed to compute the best focused point and its corresponding depth. The proposed method is experimented using synthetic and real image sequences. The evaluation is gauged on the basis of unimodality and monotonicity of the focus curve. Experimental results have demonstrated the effectiveness of the new method.

Monitoring of agent diffusion within tooth tissues is important in a wide context of tooth therapy (diffusion of medicinal preparations) and cosmetics (chemical whitening agents). We report here the results of optical coherence tomography (OCT) monitoring of diffusion of water and glycerol as clearing agents in samples of human tooth tissue. The diffusion process is analyzed by monitoring the changes in the OCT signal slope and the depth-resolved amplitude of OCT signal from a sample. Slow temporal kinetics of the mean attenuation coefficient was measured to monitor a saturable optical clearing due to the diffusion of the agent. The average permeability coefficient was estimated by dividing the measured thickness of the selected region by the time it took for the agent to diffuse through. The experimental results demonstrate that OCT can be an efficient tool in the study of agent diffusion through hard tissues.

Image Jittering was explored as a way of enhancing the peripheral vision of the AMD-Impaired. In the experiments conducted we studied the visual response of 4 normal subjects to the jittering of a grating image. Results of this study indicate an enhancement in angular resolution of about 40% (14 cpd Vs. 10 cpd) upon image jittering with amplitudes varying between 5 and 10 pixels and temporal frequency varying between 0 to 15 Hz. No significant difference in enhancement was found between orthogonal and parallel jittering motion relative to the grating direction. There was also no significant difference between the 5 and 10 pixel amplitude used. These result suggest that image jittering can be used to enhance the impaired vision of AMD inflicted patients.

The method about how to distinguish the object from the standing wave, eliminate the wave patterns in CW THz images based on phase analysis is reported. The standing wave of CW THz images result from the interference of the wave-front reflected from the object and other optics in the system. Therefore, the information of the surface of the object and the background is included in the interference pattern of the images. Because the objects have different Phase Gradients from the background, they could be extracted from the image by utilizing the phase information of the interferential waves. The waves of the image can be removed with Election-pass filter, and the result of the filtering would not deteriorate the result because there is no object in the image. At last, the image of object and the image without standing wave can be fused together to make the image better for eyes. The scope of CW THz imaging's application can be extended and the quality of images can be improved by applying this approach.