Writer verification is usually conducted by checking similarity between same type characters written by known and unknown writers. However, in the case where the same type characters do not exist in each writer’s documents, writer verification is very difficult. In this paper, we propose a method to extract the handwriting features independent of character types to solve this problem. The proposed model is based on AutoEncoder, and applying Adaptive Batch Normalization (AdaBN) and Adaptive Instance Normalization (AdaIN) for each layer of Encoder or Decoder to extract the objective features. We conducted a writer verification experiment using handwriting images pairs between different character types of ETL-1 Character Database (ETL-1). As a result of the experiment, we confirmed that the proposed method could perform writer verification with high accuracy even in such a case.
In this paper, we propose convolutional neural networks for semantic segmentation on road markings in the situation where sequential segmentation ground truth masks are available. The proposed model aggregates the temporal information and the context information from the multiple frames. Moreover, we employ CGNet as the backbone network to reduce trainable parameters and computation speed. In the experiment, we evaluate the model using the Gifu-city Road Marking Segmentation Dataset, which includes road markings of open roads in Gifu city. As a result, the segmentation performance such as a white center line and white dash line is an improvement.
The inspection of solder joints on printed circuit boards is a difficult task because defects inside the joints cannot be observed directly. In addition, because anomalous samples are rarely obtained in a general anomaly detection situation, many methods use only normal samples in the learning phase. However, sometimes a small number of anomalous samples are available for learning. We propose a method to improve performance using a small number of anomalous samples for training in such situations. Specifically, our proposal is an anomaly detection method using an adversarial autoencoder (AAE) and Hotelling’s T-squared distribution. First, the AAE learns features of the solder joint following the standard Gaussian distribution from a large number of normal samples and a small number of anomalous samples. Then, the anomaly score of a solder joint is calculated by Hotelling’s T-squared method from the features learned by the AAE. Finally, anomaly detection is performed by thresholding using this anomaly score. In experiments, we show that our method performs anomaly detection with few false positives in such situations. Moreover, we confirmed that our method outperforms the conventional method using handcrafted features and a one-class support vector machine.
Designing the optimal architecture of neural networks is an important issue. However, since this is difficult even for experienced experts, automatic optimization of the network architecture is required. In this study, we regard this issue as a combinatorial optimization problem, and utilize genetic algorithm to optimize the network architecture. Because training the networks, which are represented by individuals in GA, takes a long time, a novel method to reduce the training time by inheriting the weights of the trained network is proposed. From experimental results, our proposed method achieved the time reduction and higher accuracy than a conventional method.
We propose a defect detection method of solders on a printed circuit board using X-ray CT inspection system and Adversarial Autoencoder (AAE)[1] . We obtain sliced images of the solder using X-ray CT and extract their features that follow the standard normal distribution by using AAE. Then, the solder defects are detected by Hotelling's T square[2]. As a result of experiments, we show that we can classify normal and anomalous data samples completely on the condition of training with large normal samples and small anomalous samples.
Writer identification is one of the active areas of research. It is important to prepare a large number of characters of the same class to improve the accuracy of writer identification. However, it is not always possible to prepare enough characters of the same class. In this case, handwriting examiners compare different classes of characters and analyze using common handwriting parts for each character. However, this is very difficult. Therefore, we assume that handwriting characters written by the same writer have features independent of character classes. In this paper, we propose methods to extract features that are independent of character classes using deep neural networks. We used Conditional Variational AutoEncoder (CVAE) as a learning method. A writer identification experiment shows that these methods can extract independent features of character classes, and extracted features are useful in writer identification. Furthermore, we examined the relationship between human interpretation of character features and accuracy of writer identification by using character features extracted by disentangled feature extraction methods.
In this paper, we aimed at discrimination of defects under conditions where there is a large number of good products and a small number of defective products. Although automation of a visual inspection is essential to improve the quality of products, either or both of the features extracted by the experts and balanced dataset are needed. We tackled such a problem. By combining AAE, which can extract features following any distribution and Hotelling's T-Square, which is an effective anomaly detection method when data follows a normal distribution, it is possible to discriminate defects with a small number of defective samples.
This paper describes an automatic detection method for small dents on a metal plate. We adopted the photometric stereo as a three-dimensional measurement method, which has advantages in terms of low cost and short measurement time. In addition, a high precision measurement system was realized by using an 18bit camera. Furthermore, the small dent on the surface of the metal plate is detected by the inner product of the measured normal vectors using photometric stereo. Finally, the effectiveness of our method was confirmed by detection experiments.
In this paper, we propose a measurement system to evaluate the swallowing by estimating the movement of the thyroid cartilage. We developed a measurement system based on the vision sensor in order to achieve the noncontact and non-invasive sensor. The movement of the subject’s thyroid cartilage is tracked by the three dimensional information of the surface of the skin measured by the photometric stereo. We constructed a camera system that uses near-IR light sources and three camera sensors. We conformed the effectiveness of the proposed system by experiments.
The purpose of this study is to construct a model of an optical lens by using a multi camera array. It is known that
virtually focused images can be produced by synthetic aperture focusing techniques. However there is a difference
between the blur of the virtually focused image and the blur of an image produced by an optical lens. We suggest a method
to correct this difference. Using our method, it is possible to create images that have multiple discrete focus depths,
something that is impossible using an optical lens. Basic experiments were conducted, and the effectiveness of the
approach was demonstrated.
Nowadays, there are many instant powdered soups around us. When we make instant powdered soup,
sometimes we cannot dissolve powders perfectly. Food manufacturers want to improve this problem in order
to make better products. Therefore, they have to measure the state and volume of un-dissolved powders.
Earlier methods for analyzing removed the un-dissolved powders from the container, the state of the
un-dissolved power was changed. Our research using ultrasonographic image can measure the state of
un-dissolved powders with no change by taking cross sections of the soup. We then make 3D soup model from
these cross sections of soup. Therefore we can observe the inside of soup that we do not have ever seen. We
construct accurate 3D model. We can visualize the state and volume of un-dissolved powders with analyzing
the 3D soup models.
The components of the food related to the "deliciousness" are usually evaluated by componential analysis. The
component content and type of components in the food are determined by this analysis. However, componential analysis
is not able to analyze measurements in detail, and the measurement is time consuming. We propose a method to measure
the two-dimensional distribution of the component in food using a near infrared ray (IR) image. The advantage of our
method is to be able to visualize the invisible components. Many components in food have characteristics such as
absorption and reflection of light in the IR range. The component content is measured using subtraction between two
wavelengths of near IR light. In this paper, we describe a method to measure the component of food using near IR image
processing, and we show an application to visualize the saccharose in the pumpkin.
Nowadays, many evaluation methods for food industry by using image processing are proposed. These methods are
becoming new evaluation method besides the sensory test and the solid-state measurement that have been used for the
quality evaluation recently. The goal of our research is structure evaluation of sponge cake by using the image
processing.
In this paper, we propose a feature extraction method of the bobble structure in the sponge cake. Analysis of the bubble
structure is one of the important properties to understand characteristics of the cake from the image. In order to take the
cake image, first we cut cakes and measured that's surface by using the CIS scanner, because the depth of field of this
type scanner is very shallow. Therefore the bubble region of the surface has low gray scale value, and it has a feature
that is blur. We extracted bubble regions from the surface images based on these features. The input image is binarized,
and the feature of bubble is extracted by the morphology analysis.
In order to evaluate the result of feature extraction, we compared correlation with "Size of the bubble" of the sensory
test result. From a result, the bubble extraction by using morphology analysis gives good correlation. It is shown that
our method is as well as the subjectivity evaluation.
This paper proposes a method to detect moving objects by the background subtraction using the normalized correlation
matching. The normalized correlation matching is known as one of general-purposed template matching methods. And
the method is robust against change of brightness. Therefore, it is expected that the stable detection of moving objects
will be performed by using the normalized correlation matching against changing brightness of background. The
proposed method regards the background image as the template image and evaluates correlation rates between the
background image and the scene image in order to extract moving objects. We also adopt the integration technique of the
correlation rate to realize more stable detection.
Now a day, many evaluation methods for the food industry by using image processing are proposed. These methods are becoming new evaluation method besides the sensory test and the solid-state measurement that are using for the quality evaluation. An advantage of the image processing is to be able to evaluate objectively. The goal of our research is structure evaluation of sponge cake by using image processing. In this paper, we propose a feature extraction method of the bobble structure in the sponge cake. Analysis of the bubble structure is one of the important properties to understand characteristics of the cake from the image. In order to take the cake image, first we cut cakes and measured that's surface by using the CIS scanner. Because the depth of field of this type scanner is very shallow, the bubble region of the surface has low gray scale values, and it has a feature that is blur. We extracted bubble regions from the surface images based on these features. First, input image is binarized, and the feature of bubble is extracted by the morphology analysis. In order to evaluate the result of feature extraction, we compared correlation with "Size of the bubble" of the sensory test result. From a result, the bubble extraction by using morphology analysis gives good correlation. It is shown that our method is as well as the subjectivity evaluation.
In recent years, researches of the facial part acquisition system for the automobile driver support has been made actively. So we are developing a drivers' support system which uses a camera instead of touching with the drivers. In this paper, we use a special photography method to remove the background which disturbs the facial part acquisition. And we propose the new method by which only the face region of a driver is stably obtained in the car. Further, the eye region is detected by using the obtained facial region image in order to apply to the detection of the drowsiness.
In this paper, we propose a new Hough transform algorithm, Least Median of Squares (LMedS) Hough transform, which uses the measure of the least median of squares as the basis to estimate lines. This means that LMedS Hough transform can provide a new measure for finding lines as an alternative to the majority standard of the ordinary Hough transform and, therefore, that LMedS Hough transform can detect lines in the same way as LMedS line fitting procedure. In addition to this, because this algorithm is constructed on the Hough transform paradigm, the basic properties of Hough transform such as noise robustness, multi-line detection and global line detection are inherited in LMedS Hough transform algorithm.
Nowadays, the image processing techniques are while applying to the food industry in many situations. The most of these researches are applications for the quality control in plants, and there are hardly any cases of measuring the 'taste'. We are developing the measuring system of the deliciousness by using the image sensing. In this paper, we propose the estimation method of the deliciousness of a sponge cake. Considering about the deliciousness of the sponge cake, if the size of the bubbles on the surface is small and the number of them is large, then it is defined that the deliciousness of the sponge cake is better in the field of the food science. We proposed a method of detection bubbles in the surface of the sectional sponge cake automatically by using 3-D image processing. By the statistical information of these detected bubbles based on the food science, the deliciousness is estimated.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.