In the diffraction grating used as holographic optical element (HOE), it not only needs a high diffraction efficiency but also excellent wavelength and the incident angle characteristics over a wide range are required. We analyze the electromagnetic wave scattering problem from multilayer-coated Fourier grating for a general angle of incidence and arbitrary polarization. This analysis is treated a quasi-two-dimensional problem in the scalar boundary value. This formulation is also useful expression because R-matrix propagation algorithm is avoided a singularity of matrix elements for the evanescent mode. Numerical examples are presented for diffraction efficiencies which incident angle and azimuth angle of Fourier grating are variable.

View full abstract

Currently, 8x8 integer transformation was added to H.264/AVC for enhancing the performance of high profile, while much more complexity increased. All Zero Block (AZB) detection algorithms can help to early terminate DCT and quantization parts to reduce complexity with almost no PSNR loss. Due to the great complexity of 8x8 integer transform and quantization, however, the previous works are only concentrated on the 4x4 SAD all zero block detection. This paper proposes SAD and SATD based 8x8 AZB detection algorithms: By using the statistical analysis instead of complicated matrix derivation, through 17 QCIF sequences, 12 CIF sequences and 7 sequences of 720p test, which covers most video application, the 8x8 SAD value based all zero block threshold has been calculated. While using the SATD as the block matching criterion, the relationship between the 8x8 SAD and SATD has been derived theoretically from the energy point of view. Finally, according to this relationship and proposed SAD thresholds, the 8x8 SATD AZB thresholds have been derived. The proposed SAD value based AZB detection algorithm has saved nearly from 5% to 40% DCT and quantization calculation for QCIF, CIF and 720p sequences, while the SATD value based AZB detection algorithm achieved about from 10% to 43% calculation reduction.

View full abstract

In this paper, a new design scheme of multi-dimensional linear-phase lapped orthogonal transform (LOT) for image and audio coding is proposed. For even N, LOT basis functions described by (N × 2N) matrix, can be generated by rotation and permutation of Nth-order normalized orthogonal vectors in 2N-dimensional orthogonal space. These operations generate rotation groups and permutation groups and by using properties of these groups, the number of parameters which optimize coding gain can be reduced effectively. Furthermore several design examples of LOT are given to validate the proposed scheme.

View full abstract

We propose a 3D visualization technique which realizes browsing of hierarchical or metadata-based link structures intelligibly while avoiding 3D cluttering. This paper describes the algorithm which relocates the metaphor of contents mapped onto a 3D space so that cluttering is avoided on a 2D projection space by applying the least squares method. Also, we evaluate the effect of avoidance of cluttering using proposed technique.

View full abstract

In this paper we propose a method of synchronizing automatically the full-view image, inertial sensor and GPS information which are recorded independently, respectively, for route map building. The full-view image sensor is composed of a pair of fisheye cameras. By synchronizing the two image streams of the pair of fisheye cameras a seamless full-view image is acquired. By synchronizing the inertial sensor with the full-view image the inclination caused by route situation, such as slope and slant, can be compensated. By synchronizing the full-view image with GPS information the full-view image stream is re-sampled uniformly so that the full-view image is registered to a digital map with approximately equal distance along routes. The above synchronization is achieved by matching the motion of these sensors. To estimate the motion between consecutive frames, a small motion model of spherical image is used. The experimental results show the effectiveness of our method.

View full abstract

In this study, a method for tracking a teacher and recording teaching scene automatically during class is proposed. When a teacher walks in a classroom, the walking trajectory of the teacher is non-linear, the distribution of the motion state is non-Gaussian and the target is non-rigid. Moreover, the representation of the teacher is constantly changing; these changes include the change in the shape and color of the representation, attributed to the fact that the teacher may turn around during class. In order to solve these visual tracking problems, two images are captured, one taken when the teacher faces the classroom and the other taken when the teacher faces the blackboard to develop the color reference histogram models. During tracking process, the tracker selects an approximate reference color histogram model on the basis of the direction of the face of the teacher to obtain the accurate tracking result. The particle filter algorithm is used to choose a set of samples (particles) and let their mean distributions approximate to the target. The experimental result shows that the proposed tracking method is stable and sufficiently effective for tracking teacher, even when the teacher turns around during class.

View full abstract

Recently, it is pointed out that there exists some vulnerabilities and threats in biometric systems. In this paper, we focus on the vulnerability called ‘estimation’. This vulnerability causes a threat of impersonation by the estimated biometric information obtained through the hill-climbing attack, and another threat of impersonation by reconstructed biometric information obtained through the target user's biometric template stored in a biometric system. In this paper, we discuss the vulnerability of estimation, the threats of using it, and its countermeasure in a writer recognition system.

View full abstract

The personal computer using a two-dimensional fast Fourier transform calculation of the time, a very fast way to calculate the Kinoform, The body count of 512 × 512 or 262, 144 high-speed processing of large amounts of data to points, the image can be tone expression, a simple Kinoform can create an interference pattern faster only CPU. Have been considered the table in the long times. However, using the GPU or multi-processor is a general-purpose computation, such as increasing attention is the high processing power. This method was, CPU or GPU using to compare the operating time of Computer-Generated holography of real 3D object taken by Web-cameras is considered in the future.

View full abstract

[in Japanese]

View full abstract