In this paper, we propose a new image retrieval method based on binary local features and Bag-of-Visual Words framework. We extract uncorrelated portion from bit strings of feature vectors in the same cluster in order to efficiently represent the distance between the vectors. The effectiveness of the proposed method is confirmed in experiments.

抄録全体を表示

In this paper, we proposed a video segmentation method during field operations. The proposed method discriminates scene differences based on operator's gaze and viewpoints which can be estimated through equipment detection against operator. Experiments revealed that continuities in time domain are important for our clustering procedures to obtain suitable segmentation results.

抄録全体を表示

We present a fast method to solve discrete labeling problems based on cost volume filtering. By region splitting and voting, the number of sub-regions in cost slices are reduced. As a result, our method can reduce computational time to 60%〜80% of the original algorithm, while keeping the accuracy.

抄録全体を表示

We evaluate the performance of the proposed stable computation scheme for the variational scene-flow estimation, whose stability is independent of the regularization parameter and stereo image sequences. The results show that the proposed method gives smaller endpoint error than the previous method.

抄録全体を表示

When we shoot a particular person from a fixed position, it is difficult to catch the person always in frames due to occlusion and so on. This paper proposes a method to automatically extract scenes that particular person is reflected from videos shot from multiple directions while tracking other person.

抄録全体を表示

Smile is a facial expression that is expressed when the emotion of joy came out. Smile to express different feelings. However, it is classified in the same category but different meaning. So I were classified into three categories the expression of the smile.

抄録全体を表示

In this study, we propose a mathematical model of the retinal network considering the characteristics of peripheral vision to evaluate the influence of fixation eye movements on the visual perception. Simulation results show that microsaccades enhance the responses of M-type ganglion cells for a vertically sinusoidal gratings, but maintain their resting state for a random dot pattern which consists of a particular spacial frequency.

抄録全体を表示

The saliency based mathematical models are frequently used to predict observer's gaze shifts. However, evaluation methods of scan path for several gaze steps do not exist. In this study, we proposed a new index by which the similarity of feature vectors on the passed gazed locations were veraged. The results suggest that our index is effective to compare the similarity of simulated scan path and human eye movements.

抄録全体を表示

In this report, we examine effects of stereoscopic vision, olfaction, and haptic sense on the sense of presence in a virtual environment.

抄録全体を表示

We investigated the effect of motion parallax using natural scene images. At first, we examined the optical flow in natural scene images, when the capture device moved horizontally and sight on the fixation point. Second, we pick up the natural scene using that kind of motion capture device. For these captured natural scene images, we evaluated the effect of motion parallax.

抄録全体を表示

In this paper, we report the results that analyzed relationship between Luminance of View Image and Pupil Diameter.

抄録全体を表示

A cause of accidents of elderly drivers is due to impaired judgment and caution and the lack of a wareness. In the present study, we developed a method for measuring and evaluating the deterioration of elderly cognitive function while driving using a new driving simulator.

抄録全体を表示

Since novice drivers cause a serious accident more than experienced drivers, the system for supporting the reeducation to novice drivers is demanded strongly. In this study, we develop a method to measure for predicting accident risk with a driving simulator.

抄録全体を表示

This study is intended to examine of effectiveness of using video in business operations. First, clarify the haracteristics of video from previous researches. Then, clarify the effectiveness to use video in this field by the case studies and discuss about the possibility of the future.

抄録全体を表示

A mobile shape recognition system using the histogram of Radon transform (HRT) is proposed. Our HRT descriptor based on the Radon transform gets robustness to shapes rotation, scaling, and translation. Our system is performed and we confirm the recognition performance.

抄録全体を表示

We describe a "size bias" effect leading to tracking errors in a face recognition system based on multi-resolution deformable template matching. Countermeasures are proposed and shown to be effective with experimental data.

抄録全体を表示

We present a method to extract incorrect recognition results for object recognition by analyzing the confidence value yielded from the machine learning algorithm in detail. In this paper, we demonstrate that the difference between the confidence for the 1st candidate and that for the second candidate is a good measure for such true-false judgment. Experimental results show that our proposed algorithm is about 50% better than the conventional simple thresholding using the confidence value.

抄録全体を表示

We have developed the super-resolution technique for scanning the high density bar-code that the camera image is distorted. It was impossible to apply to heavily tilted bar-codes. We propose the method based on the image partitioning to realize the super-resolution even if the bar-code is tilted at any angle.

抄録全体を表示

Slanted-edge method is well known as the international standard for measurement of the Modulation Transfer Function(MTF). It is, however, limited to the case only if angles of the edge to be measured are small. In this paper, we propose the method to measure the MTF accurately even if the angles are large.

抄録全体を表示

In multi-monitor environment, we examined the playback control method for the asynchronous playback of different content, and the synchronized playback of the same content.

抄録全体を表示

Smart devices like smartphones can only individually play Audio and Visual(AV) contents. But smart devices can't synchronously play the AV content between devices. We propose an AV Synchronized Playback System that is able to play an AV contents synchronously between smart devices.

抄録全体を表示

We propose the accelerating channel selection processing method which makes it possible to display a video earlier than conventional channel selection processing method when you select the TV channel.

抄録全体を表示

We propose the fast boot-up processing method which makes it possible to show the first video frame faster than conventional methods when you turn on the TV.

抄録全体を表示

We have conducted experimental studies on VNC (Virtual Network Computing) use to transfer a PC screen to a TV display. Because of the limitation of CPUs performance implemented in a TV, the screen on TV sometimes changes slowly. In this paper, we will describe our experimental results concerning transition models (speed, changing shape) of the screen which requires maintaining comfortability level for visuals.

抄録全体を表示

We have developed a software interface using gesture recognition technology, which enables rotating CG character model for virtual studio use. Usability has been examined and shown to be easy to use.

抄録全体を表示

This technique detects partitioning and the layout of the room using an image photographed with the camera of the air-conditioner. And this technique controls the direction of the wind and the air capacity of the air-conditioner.

抄録全体を表示

We propose a synchronization of the music with playing a musical instrument and automatic accompaniment system. The proposed method estimates the tempo from the motion data and musical signal in musical performance and we use Kinect sensor to retrieve the motion data.

抄録全体を表示

There has been an efficient planar metric rectification method for a multiple-plane environment when the angle of the planes is given. However, the number of boundary images used for the angle constraint is limited to only one. This paper proposes a high-precision method with multiple boundary images.

抄録全体を表示

This paper represents a degradation of real objects based on 3D models that made from the RGB-D camera. We obtain the depth information from the RGB-D camera, and create 3D models of real objects. We apply degradation map to 3D models, and superimpose 3D models to real objects.

抄録全体を表示

The purpose of this study is to construct metallic texture space for understanding the space easily as a color space. The results of subjective evaluation are analyzed using principal component analysis for deciding the axis of the metallic texture space.

抄録全体を表示

In the study of Augmented Reality, an object is desired to generate a new object image with different colors from an original image but do not alter the gloss of the object. This research proposes a method of color alterati on of real object using Augmented Reality.

抄録全体を表示

We are developing a bio-implantable glucose-monitoring device based on complementary metal-oxide-semiconductor (CMOS) technology. We investigated basic measuring capability of CMOS image sensor with glucose-sensitive fluorescent hydrogel. This result suggest that our device has capability of measuring glucose concentration.

抄録全体を表示

We developed a high speed and low power tdcSS AD converter applying TDC technology using "the asymmetry type oscillator" to Single Slope type AD conversion technology. We achieved 1.728GHz(/Step) operation with the same low power as a conventional Single Slope AD converter in 108MHz(/Step) operation by this technology.

抄録全体を表示

We have studied a 3D-integrated image sensor that is intended to provide both an ultrahigh-definition and a high frame frequency. For such new sensors, we have developed a fundamental technology to directly bond substrates and to construct 3D-integrated logic circuits.

抄録全体を表示

Crystalline selenium (c-Se), which shows a high spectral sensitivity in the visible light region, is suitable for a photosensitive layer of imaging devices. A c-Se/ZnO p-n heterojunction photodiode has been fabricated to demonstrate the high quantum efficiency in the visible light region and low dark current under high electric field.

抄録全体を表示

We have developed a new method to extract the density of trap states from oxide thin-film transistors (TFTs) by combining simulation using a computationally efficient model and measurement by the low-frequency capacitance method. The developed method enables fast and accurate characteristic analyses for oxide-TFTs.

抄録全体を表示

Optical compensation of flexible vertical-aligned LCDs was achieved for wide viewing angle. Optical simulation clarified that viewing angle range can be improved by optical compensation considering optical anisotropy of plastic substrates. We then obtained good contrast ratio of 650:1 and wide viewing angle of 160° by fabricating devices.

抄録全体を表示

In this paper, we propose objective quality estimation method based on Parametric NR framework from H.265/HEVC bitstreams. H.265/HEVC utilizes an adaptive block size based on quad-tree structure, and any existing Parametric NR framework designed for H.264/AVC cannot be applied. Therefore, the proposed method extracts spatial and temporal features for each block size in order to estimate the picture quality. Experimental results show that a correlation coefficient between subjective quality score and estimated quality values is 0.941.

抄録全体を表示

We measure the modulation transfer function (MTF) of a Super Hi-Vision camera with the slanted-edge method. We show that the measured MTF corresponds to the contrast transfer function with inmega cycle chart method. We confirm the measured MTF of an 8K camera according to the specification similar to HD cameras.

抄録全体を表示

A modified slanted-edge method is presented for measuring the modulation transfer function (MTF) of digital image acquisition devices. We measure the MTF of an 8K camera using the modified method.

抄録全体を表示

We report a study on a spectral sensitivity correction method that uses an optimized color conversion filter to obtain a spectral sensitivity that is closer to the ideal sensitivity suitable for wide-color-gamut imaging. By optimizing the amplitudes of R/G/B, coefficients of the linear matrix, and spectral transmission factor of the color conversion filter, we show that the spectral sensitivity approaches the ideal spectral sensitivity and that the color difference decreases.

抄録全体を表示

We developed an ultra-compact Super Hi-Vision camera with 33-megapixel single-chip color CMOS image sensor. The camera head body size is as small as 125 x 125 x 150 mm and the weight is 2kg. The optical size of the image sensor is 1.5 inch and the camera head can be used with various super 35 PL mount lenses for digital cinema.

抄録全体を表示

We developed a new Super Hi-Vision (SHV) camera toward shooting theater contents. We designed signal processing for achieving higher sensitivity than the conventional SHV cameras and housing to reduce acoustic noises. We measured sensitivity and noises, and then confirmed this camera specification was satisfactory for shooting heater contents.

抄録全体を表示

We developed elliptical horn antennas with narrow vertical directionality for 42-GHz band wireless camera systems. From the results of propagation experiments on a golf hole, we found that these antennas increased average received power by 2.5 dB compared to conical horn antennas and had possibility to expand transmission ranges.

抄録全体を表示