This study explores the effectiveness of wavelet analysis techniques on digital holograms of real-world 3D objects.
Stationary and discrete wavelet transform techniques have been applied for noise reduction and compared. Noise
is a common problem in image analysis and successful reduction of noise without degradation of content is
difficult to achieve. These wavelet transform denoising techniques are contrasted with traditional noise reduction
techniques; mean filtering, median filtering, Fourier filtering. The different approaches are compared in terms of
speckle reduction, edge preservation and resolution preservation.
The nature of digital hologram's allows for the implementation of segmentation process' using volumes of reconstructions
as their input, where each reconstruction in the volume is a reconstruction at a different focal plane.
Our segmentation technique utilizes extracted focus and shape information. In the case of digital hologram's
encoding macroscopic objects, this information is generally obtained using data extraction algorithms applied to
a volume of reconstructions. We have developed a three stage segmentation algorithm for macroscopic objects
encoded in digital hologram's. This algorithm uses a depth-from-focus technique applied to a set of numerical
reconstructions from a single perspective of the digital hologram to extract focus and shape information in the
form of a depth map and a maximum focus map. First we estimate the degree of focus at each coordinate. We
then calculate a depth map of the scene and segment all object coordinates from background coordinates in the
second stage. Finally in the third stage, we perform the segmentation of a digital hologram's reconstruction into
independent objects or object regions. A segmentation image is created through applying histogram and image
processing algorithms to the depth map. We present results of applying our technique to digital hologram's
containing single and multiple macroscopic objects.
Speckle is an inherent characteristic of coherent imaging systems. Often, as in the case of Ultrasound, Synthetic Aperture
Radar, Laser Imaging and Holography, speckle is a source of noise and degrades the reconstructed image. Various
methods exist for the removal of speckle in such images. One method, which has received attention for the removal of
speckle from coherent imaging, is to use a temporally incoherent source. We create a novel digital signal processing
technique for the reduction of speckle from digital holograms by simulating temporal incoherence during the digital
reconstruction process. The method makes use of the discrete implementation of the Fresnel Transform, which calculates
the reconstructed image for a range of different wavelengths. These different spectral components can be weighted to
suit a temporally incoherent source and the intensities from each wavelength are added together. The method is
examined using the speckle index metric.
We investigate the principles of digital holography based on the Wigner distribution function (WDF). We apply the WDF to the analysis of generic optical setups which are used to record and reconstruct image Fresnel holograms. We use the graphical representation of the Wigner chart to derive various important properties, including the required space-bandwidth product of the digital hologram, CCD sampling and numerical reconstruction and the optimum required object size to optimize the system efficiency. This allows us to offer a simple comparison of the various recording schemes. The analysis also allows us to graphically compare the numerical reconstruction methods and the restrictions it may impose on the CCD parameters. We show how this insightful analysis leads us to a new method of digital holography which we call 'dual-shift' DH.
In this paper, we analysis the effect of partial occlusions in scenes captured using digital holography. We reconstruct the scene from different perspectives. These reconstructions are then combined, allowing one to overcome foreground occlusions that are obscuring one's view of the scene. The analysis in this paper is carried out with the aid of the Wigner distribution function, allowing us to visualize the energy of the object wavefield and the occluding object wavefield in phase space. We show that by iteratively selecting different views, the original scene can be reconstructed e±ciently. This technique would be useful in situations where transmission of the whole digital hologram, or exhaustive reconstruction of every perspective, was not feasible. We provide results using optically captured digital holograms of real-world objects, and simulated occlusions.
In this paper we concern ourselves with the subject of superresolution in Digital Holography (DH), i.e. increasing the resolution of DH system beyond its limit. The limiting factor regarding resolution in a DH system is the pixel size, which is equal to the smallest resolvable unit. By careful superposition of different digital holograms captured of the same 3-D object, we attempt to increase the resolution of the reconstructed image and equivalently to increase the range of angles of reconstruction. This is accomplished by rotating the input object wavefield either by rotation of the object (it is 2-D) or by rotation of a mirror that is placed between the object and the CCD. Rotating the input wavefield shifts the wavefield in the hologram plane in space and spatial frequency. Therefore, those parts of the hologram field that contained energy at too great an angle for recording and were therefore arranged to be adjacent to and not on the CCD will be shifted in space onto the CCD face and will also be shifted to a recordable angle. We outline a sub-pixel correlation technique to stitch the consecutive holograms together in both the space and spatial frequency domains. Multiple captures enable us to record a DH of large resolution and angle of reconstruction. Storage and reconstruction of the stitched hologram is also discussed and experimental results are given. The method may be applied with any existing form of DH. We use the Wigner Distribution Function to qualify and quantify the method.
We investigate the application of Independent Component Analysis to the reduction of speckle in reconstructions from digital holograms. Independent Component Analysis computes a linear transformation of a multidimensional distribution that minimizes the statistical dependence between components. It can be seen as an extension of Principal Component Analysis where the transformed bases do not need to be orthonormal. We attempt speckle reduction across multiple hologram reconstructions. A number of situations are investigated,
including recording two holograms over the interval of a day, changing the illumination between two holograms and adding a diffuser in the path of the object beam between subsequent hologram captures. This ensured significant speckle differences between the observations. Results are provided using simulated and optical data.
We propose a task-specific digital holographic capture system for three-dimensional scenes, which can reduce the amount of data sent from the camera system to the receiver, and can effectively reconstruct partially occluded objects. The system requires knowledge of the object of interest, but it does not require a priori knowledge of either the occlusion, or the distance the object is from the camera. Subwindows of the camera-plane Fresnel field are digitally propagated to reveal different perspectives of the scene, and these are combined to overcome the unknown foreground occlusions. We demonstrate that careful combination of reconstructions from subwindows can reveal features not apparent in a reconstruction from the whole hologram. We provide results using optically captured digital holograms of real-world objects, and simulated occlusions.
We report on recent advances made in the area of holographic image processing of three-dimensional (3D) objects. In particular, we look at developments made in the areas of encryption, compression, noise removal, and 3D shape extraction. Results are provided using simulated objects and real-world 3D objects captured using phase- shift digital holography.
We have successfully applied Independent Component Analysis to the removal of background speckle noise from digital holograms. Additive noise removal techniques do not perform well on speckle, which is better characterized as multiplicative noise. In addition, speckle contains 3D information and so cannot be removed completely. We use a blind source separation approach to the reduction of speckle noise in digital holograms. Independent Component Analysis computes a linear transformation of a multi-dimensional distribution that minimizes the statistical dependence between the components. It can be seen as an extension of principal component analysis where the transformed bases do not need to be orthonormal. Although a linear technique, we show how Independent Component Analysis can be applied to the reduction of background speckle in digital holograms. We have captured our digital holograms of three-dimensional objects using phase-shift digital interferometry. In addition, the technique can be extended and applied to segmentation and pattern recognition problems on digital holograms of three-dimensional objects. Results are provided using simulated and optical data.
We present a technique to convert a digital hologram of a three-dimensional (3D) object into a cloud of surface points in 3D space. Two depth-from-defocus techniques are used to generate a depth map for a particular reconstructed perspective of the object encoded in the digital hologram. The Fresnel transform is used to effect defocus, and a histogram-based approach is used to determine the degree of defocus for each neighborhood of pixels. Our experiments involve simulated and real-world objects (captured using phase-shift digital interferometry). The technique could be used in registration and 3D object recognition applications.
This paper details a procedure for classifying facial expressions. This is a growing and relatively new type of problem within computer vision. One of the fundamental problems when classifying facial expressions in previous approaches is the lack of a consistent method of measuring expression.
This paper solves this problem by the computation of the Facial Expression Shape Model (FESM). This statistical model of facial expression is based on an anatomical analysis of facial expression called the Facial Action Coding System (FACS). We use the term Action Unit (AU) to describe a movement of one or more muscles of the face and all expressions can be described using the AU's described by FACS.
The shape model is calculated by marking the face with 122 landmark points. We use Principal Component Analysis (PCA) to analyse how the landmark points move with respect to each other and to lower the dimensionality of the problem. Using the FESM in conjunction with Support Vector Machines (SVM) we classify facial expressions. SVMs are a powerful machine learning technique based on optimisation theory.
This project is largely concerned with statistical models, machine learning techniques and psychological tools used in the classification of facial expression. This holistic approach to expression classification provides a means for a level of interaction with a computer that is a significant step forward in human-computer interaction.
One of the principal successes of computer vision over the past thirty years has been the development of robust techniques for the estimation of the structure of a 3D scene given multiple views of that scene. Holography is an established technique for recording and reconstructing real-world 3D objects. A single hologram encodes multiple perspectives of the scene simultaneously, and hence provides a novel avenue of extension of these traditional computer vision techniques. In this paper, we explore the pontential use of digital holograms in 3D scene reconstruction where particular regions of interest are occluded under particular views. In our experiments we employ both synthetic holograms of artificial scenes, and optically-captured digital holograms of real-world objects. We show that by selecting a particular set of perspectives, determined by the occlusions present in the scene, the original scene can be reconstructed.
We present the results of applying data compression techniques to encrypted three-dimensional (3D) objects. The objects are captured using phase-shift digital holography and encrypted using a random phase mask in the Fresnel domain. Both the amplitude and the phase of our 3D objects are encrypted using this technique. The advantage of a digital representation of the optical wavefronts is that they can be processed and transmitted using conventional means. Arbitrary views of the 3D objects are decrypted and reconstructed using digital propagation. Compression is applied to the encrypted digital holograms prior to transmission. Degradation due to lossy quantization compression is measured in the reconstruction domain. Finally, we use a speedup metric to validate that our compression techniques are viable for time-critical 3D imaging applications. Our techniques are suitable for a range of secure 3D object storage and transmission applications.