By allocating saliency maps to the region of interests (ROIs), ROI-JPEG has been demonstrated to generate JPEG images with high compression rates without degrading perceived quality. However, on putting ROI-JPEG to practical use, some issues need to be resolved: degradation of the resulted image caused by an improper saliency map and difficulty in implementing the code due to inability to use the existing JPEG libraries. This paper proposes a simplified implementation of ROI-JPEG by separating the part of the ROI coding to assure the perceived level of qualities: limit of detection, acceptability limit and to allow the use of an existing JPEG library to save the resulted image. Our experiments show that regardless of the structural similarity between maps of saliency and eye fixation distributions, the output images will receive no degrading perceived quality and about 20 % lower bitrate compared with the ordinary JPEG images by setting the minimum quality factor for the quantization matrices during the ROI coding to 45. Meanwhile, setting the minimum quality factor to 25 will generate about 30% lower bitrate with acceptable perceived quality of images.

View full abstract

In order to easily examine human visual performance of colors (Color Visual Acuity: CVA) when conducting medical care in eye clinic, we have developed a new testing system that can measure CVA using PC and LCD (Liquid Crystal Display) achievable color calibration accurately. In this testing system, we utilize Landolt ring as a stimulus to be shown to subject. Creation of colors used in the test, management of data file for Landolt ring (stimulus), and operation of test can be done by using the color management and the stimulus management applications (software) implemented on PC. As for color reproduction on LCD, we checked the color difference between target (theoretical) value and measured one for each color in NEW COLOR TEST and White point D65 used in the test as well as the brightness difference between each color and White point by using high-performance spectroradiometer. The result reveals that RMSE for both color difference and brightness difference are very small so that the LCD can be available for CVA test. As an example of measurement using this testing system, we examined CVA for normal eyes of young subjects as well as CVA for eyes of aged subjects, which proves the usefulness of this testing system.

View full abstract

Recently, recognizing a human expression plays an important role in a machine cooperating with people. This paper proposes a method to recognize expression in facial images using novel features. First, facial key parts are extracted by using a generic Active Appearance Model (AAM), then Gabor magnitude pictures are obtained by convolving the images of facial key parts with Gabor filters, followed by encoding with Local Directional Pattern (LDP) operator which enhances facial local feature. The maps generated by these processes are divided into several blocks, and Local Gabor Directional Pattern Histogram Sequence (LGDPHS) based on facial key parts are extracted from each block. The histograms are concatenated to a vector, where Principal Component Analysis (PCA) is used to reduce the dimensions. Finally, the feature vector is classified by Support Vector Machine (SVM). In our experiment, we demonstrate both person-independent and person-dependent facial expression recognitions (FERs), and the recognition rate of our person-independent FER improves up to approximately 16% as compared to the conventional methods. The performance of our recognition reaches to 94.74% in person-dependent FER. As a result, we have successfully demonstrated effectiveness of our proposed method.

View full abstract

This paper investigates a dual-function face recognition home security system using a hybrid Fisher linear discriminant (HFLD) analysis method that is able to authenticate an individual who is trying to enter the premises and also identify each individual in the household. Conventional Fisher linear discriminant (FLD) is a practical, robust algorithm that is suitable for a home security system, however, input test images of a unknown face class that is not from the training gallery results in low success rates. To solve this problem, we propose a HFLD that combines the benets of both global and local approaches by dividing the face image into ve components i.e., Fisher-face, Fisher-eyes, Fisher-eyesnose, Fisher-nosemouth and Fisher-mouth. It was observed that the proposed method achieved 99:2% and 98:6% success rates in face authentication for the AT&T and MZ databases, respectively. This method also yielded success rates of 97:5% and 96:3% for face identication for the AT&T and MZ databases, respectively.

View full abstract

Demands to obtain high-resolution images are increasing recently in many appliances, such as large TV units, HD cameras, mobile phones and so on. Many interpolation methods such as bi-linear, bi-cubic, and Lanczos methods have been developed which up-scales digital images, however, they result in high resolution blur, because they do not include estimating higher frequency components. Thus, we propose an up-scaling method using both morphological operation and image separating based on the Total Variation method. The proposed algorithm reduces environmental noise as well as blurring from the result. As an experiment, we prepared a down-sampled QVGA image and then, applied the Lanczos resampling, the Laplacian pyramid, and the proposed method. The result images successfully demonstrate the effectiveness of the proposed method, while the MSE figure shows some degradation in some images.

View full abstract

Recently, virtual reality (VR) environments where virtual objects can be accurately operated and the real world is closely mimicked have been constructed using haptic devices and physics simulators. To control virtual objects in physics simulators by operating haptic devices, it is necessary to connect the haptic pointer serving as the human hand in the VR environment and the grip of the device actually operated by users. As a means of realizing this, virtual coupling, which connects the two using a spring-damper model, can be used. In virtual coupling, devices may oscillate and their fidelity may decrease depending on the coupling parameters; therefore, application developers have empirically determined the coupling parameters for each application. In this study, we propose methods of determining the optimal coupling parameters by genetic algorithm.

View full abstract

Domestic robot with voice communication aid which is controlled by human eyes only is proposed for disabled and elderly persons. It allows taking a look the surroundings of the domestic robot using cameras mounted on the domestic robot. It also allows enjoying conversations with the surrounding peoples of the domestic robot using voice output capability of the domestic robot. It can avoid obstacles autonomously with not only camera acquired images but also ultrasonic sensors. If users look at the top left and right as well as bottom left and right, then the domestic robot goes back to the previously designated home position. Such functionalities are confirmed with normal persons together with key-in success rate, blink detection accuracy, comparative study on key-in success rate and process time between fixed (conventional) and proposed moving keyboards.

View full abstract

In this paper a design of an optical-transparent wireless link with channel encoding and bit interleaving for enhanced digital TV transmission is proposed. For completing the system simple channel code is chosen and original all-optical schemes of a millisecond delay line and bit interleaving and de-interleaving modules are proposed and explained in detail. Considering the limitations of the used technologies and the overall delay of the wireless link, the example in the paper is for a single-wavelength 40Gbit/s signal. The interleaving degree is chosen to be 1.5ms based on the presented experimental data and theoretical analysis for burst error length. Simple evaluation of the system performance is provided and the results are compared to similar researches and discussed.

View full abstract

We calculate the Fourier transform of two-dimensional data from the camera input of a real three-dimensional image and create a hologram. Depending on the phase modulation, a hologram can be either a kinoform or a phase hologram. Kinoform has been studied previously. They produce a reactivation image, but matching is difficult because noise arises when using only phase information. Therefore, we propose a method to produce a phase hologram with a high precision system that creates a hologram based on interference fringes using the phasing and amplitude of the new system and displays the intensity distribution on a phasing device. The liquid crystal device tested with this system shows a degree of phase modulation of greater than 2π and a high diffraction efficiency. We compare the kinoform with a phase hologram using a high resolution liquid crystal device that modulates the phase and report that the phase hologram demonstrates better performance.

View full abstract

We propose a low-complexity pixel-based watermarking using edge area in an image. In this method, an image is divided into blocks and the blocks are classified into edge blocks and others. The histogram of pixels in edge block generally shows a bi-modal distribution which has two peaks. Pixels in an edge block are further classified into two classes named “high peak class” and “low peak class”. Then, a bit of watermark information is embedded by manipulating the between-class variance, i.e., by changing the pixel values in each class of the edge block. We utilize the discriminant analysis method for deciding a threshold and manipulating the between-class variance. The proposed method can improve the quality of watermarked image while keeping the robustness against compression and smoothing. In this paper, we evaluate the image quality of watermarked image using PSNR and SSIM. We also evaluate the watermark robustness against JPEG compression and Gaussian filtering.

View full abstract

In modern societies, formal standardization is one means to help balance the commercial desire for profit with the public desire for open standards. How does/should a standardization committee balance the importance of private gain to motivate the creation of the new and improved with the importance of open standards to expand markets and define open interfaces for future extensions? This paper presents a broad historical view of standardization and describes how different successions of standards have been balanced in quite different ways. A technical approach to balance compatibility standardization and standards is identified along with successful examples of its implementation.

View full abstract