A high-speed transcoder working with video watermarking is developed to deter information leak of relatively small size video content. Almost invisible watermarks can survive against severe image processing in video sharing services.

抄録全体を表示

We propose a tracking algorithm based on mean shift procedure for occlusion. Mean shift tracking is rubust for partial occulusion. However, this method fails to track, when the most of the part of the object is occluded. We introduce a criteria of occulusion to rearrange the window if it is observed. Experimental result shows the efficiency of the proposed algorithm.

抄録全体を表示

Although optical flow is useful for object tracking, estimation of the flow by the gradient method is ill-posed problem. In order to avoid ill-posedness, we have proposed a tracking method via 1D flow, which is calculated on the straight lines (calculation axes). We update these axes based on the positive 1D flow. Since, origin of the axes is located on the restricted position, we can't reduce a tracking error. We propose tracking method by using both 1D flow and frame difference.

抄録全体を表示

We often use video contents over the broadband network. We had proposed the CDSS: "Content Download Service System" based on human behavior as video content delivery systems. Especially, we pay attention to the cache delivery system. We consider the effect of the proposed system based on time function by simulation experiments.

抄録全体を表示

The proposed visible light communication method uses visually non-lighting LED, which enable us to communicate by using tail lamps of cars. In addition, this method can communicate with visually non-lighting traffic lights.

抄録全体を表示

As one of the mobile phone contents, electronic books and magazine, especially comics, are expected to have potential possibility. In viewing comics on mobile phones, the resolution and the size of screens must be inadequate. A comic then needs to be decomposed into frames and put them in order to be read properly, whereas it is known that an accurate decomposition is not easy. In this paper, we propose an accurate and fast clipping method using contour tracing and evaluate it comparing with a commercial technology.

抄録全体を表示

The players of badminton game need more efficient smash motion. This paper proposes a new method to analyze ideal smash motion. An experiment shows the result of having applied the proposed method to the video of Japanese elite players.

抄録全体を表示

In this paper, we consider a method to improve the precision of GPS by camera on mobile phone. In order to investigate the robustness of the proposed method for mobile phone camera properties, we apply motion pictures by various types of mobile phones for the performance evaluation.

抄録全体を表示

We designed a system for achieving automatic, stable and accurate tracking of persons walking around within the overlooked space by means of cooperative operation of both an omni-directional camera and Pan/Tilt/Zoom cameras. Stability of the tracking performance was investigated by experiments.

抄録全体を表示

This paper describes a motion recognition method for physical exercise(pommel horse) by image processing technique.

抄録全体を表示

We had developed a broadcasting quality single-chip color HDTV camera using 1.25-inch 8.9M-pixel 60fps CMOS image sensor. In this study, we tested the camera for drama shooting. With large size imager and small F-number PL mount cinema lens, this camera realized a shallow depth of field and was well suited for drama shooting.

抄録全体を表示

We have developed the new alignment method of telescope pointing for Low-earth-orbit satellite shooting system. By adopting statistical method for this system which has three axes to shoot pictures of satellites, the pointing precision has improved up to 0.1 degrees in the deviation angle.

抄録全体を表示

TWINSCAM is a camera system that synthesizes the two images shot with two cameras above the surface of the water and below the surface of the water at the side of the swim pool. And it achieved a new effect of the images in live TV programs of synchronized swimming.

抄録全体を表示

TWINS CAM was used by live broadcasting for the first time in synchronized swimming JAPANOPEN 2010. The movement below the surface of the water and above the surface of the water such as LIFT and SPIN was able to express the image that is expressible by one picture, and not is before.

抄録全体を表示

Energy harvesting is a power-generating technology from environmental energy. We propose applications of batteryless/wireless remote controller to broadcasting engineering using this technology. The prototypes remote controllers of TV camera and jack for an outside broadcast van were developed and tested in some program productions, so that their usability was confirmed.

抄録全体を表示

We propose a method for store-and-forward video data transmission with time constraints that sets encoding parameters automatically considering time required for encoding and transmitting. Our method processes the video data in the most efficient order to maximize the quality of transmitted video data.

抄録全体を表示

Authors are seeking universal services to be applied to broadcasting. As the one of them, we focused on assisting elderly viewers in audio, and built an audio processing system. The processes are based on compensating their hearing characteristics. Viewers can listen to the programs with processed audio without special devices.

抄録全体を表示

In this paper, an automatic overlay advertisement for video content is described. By considering the amount of visual feature and composition of the video frames, we realize the least intrusive advertisement overlay for video services.

抄録全体を表示

We developed a trial content in Broadcast Markup Language for data broadcasting services using an individual QR Code linked to the shelter map on a mobile phone. Effective use of the bandwidth allocated to data broadcasting services was achieved by generating QR Codes in a receiver.

抄録全体を表示

Recently, a server virtualization technology is used for various systems like web servers. However, it hasn't been applied to the broadcasting systems yet. This report is about a trial of the terrestrial digital broadcasting systems using a server virtualization technology.

抄録全体を表示

実際に放送された番組に関する評判をデータマイニング技術を用いて解析した結果を報告する。

抄録全体を表示

We developed relevant retrieval and recommendation system for NHK creative library. This system used relevant retrieval technologies and a system model of contents recommendation television that retrieves content from the selected content by metadata. The system provides a function to search for video of creative library in a joyful way.

抄録全体を表示

We studied a method for creating an image DB using image sharing sites to select images expressing atmosphere of weblogs automatically. We found that using comments for search by impression words is more effective than using image tags especially when an image has many comments including the same impression word.

抄録全体を表示

Destiny of research work is to judge similarity between two documents. We focus documents that are on computer and they are assigned by character code. We transform them by Fourier, Fourier cosign, Haar wavelet and Hadamard transforms. We calculate barometer of similarity judgment by their transform's average of cross correlation.

抄録全体を表示

We propose a retrieval method using Wikipedia revision history as an index of mass attention. Preliminary experiments show that the proposed method has the potential to improve results based on word frequency.

抄録全体を表示

We have proposed personal adaptation in TV content recommendation. This experiment is a trial to search TV content which is suitable for an audience. This paper describes an outline of the architecture for personal adaptation in TV content.

抄録全体を表示

This paper proposes a viewpoint-dependent display system which gives pseudo 3D perception based on motion parallax. The system has a fisheye camera to detect position of an observer's head. Experimental results show that the mean-shift algorithm works well for human head tracking in video sequences with large lens distortion.

抄録全体を表示

We developed interactive broadcasting system named "Nyafu Nyafu Douga". Using this system, viewer can watch broadcast and post comments at the same time. We put this system into practice in University of Tsukuba school festival.

抄録全体を表示

An electronic book market accomplishes rapid grown. Comics for mobile phone occupy more than 80% of the electronic book market in Japan. Image processing is necessary to display comics on mobile phone. In this paper, we examine statistical characteristics of comic frames for the image processing.

抄録全体を表示

This paper proposes a method of extracting typical images as templates for structuring video program. The template consists of a typical image of production style and a related mask image which eliminates changing parts of the style. The method was applied to some News programs and its effectiveness was confirmed.

抄録全体を表示

In this paper, we proposes a new adult image detection method based on frequency characteristics of optical flow vectors. And, we show that our proposal is a very powerful method for adult image detection by experimental results.

抄録全体を表示

A lot of accidents that the visually handicapped person falls to the railway track have occurred. In our research, it aims at the prevention of accidents that falls from such as the platform edge and the descending stairs. In this text, we report on the constructed research system. And, we consider the surrounding situation sensing for visually handicapped person using USB camera and ultrasonic ranging sensors.

抄録全体を表示

We investigated the effect of auditory spatial attention on brain activation in dichotic listening task. Participants paid attention toward an auditory input from their left or right ear. Cerebral hemodynamic response was measured by a functional near-infrared spectroscopy (fNIRS) in temporal lobe, occipital lobe, and posterior parietal cortex (PPC) in both lateral cerebral hemispheres. In both lateral PPCs, the attention toward input from left ear induced significantly larger activation than from right ear. Harasawa and Shioiri (2008) indicated that visuospatial attention to left visual hemifield also induced significantly larger activation in PPC than to right hemifield. These suggested that there might be asymmetrical spatial attention regulation system which was independent on sensory input modality.

抄録全体を表示

In cosmetic research, evaluation of the skin softness and elasticity has usually obtained from numerical data. However, it is difficult to understand the degrees of skin softness and elasticity from numerical data for the public. This approach was to develop the user-friendly visual expression by using high-speed camera.

抄録全体を表示

This paper studies a method of detecting roads with trees by analysing public streetscape images on the web in order to reduce the cost of constructing radiowave propagation models. While the method is simply based on color features, the detection accuracy is as high as 90 %.

抄録全体を表示

In order to remove remote control bar from a television, gesture control interface is introduced in this prototype. The most recognizable shape and movement of hands is studied and adopted in this prototype which allows users to operate it by simply moving hands.

抄録全体を表示

We propose a gesture recognition method that uses 3.5-dimentional trajectory feature descriptors. We measured horizontal, vertical and depth information from a grayscale image and a depth map by using a time-of-flight camera. The method can be applied to a hands free human interface.

抄録全体を表示

In order to make it possible to presume the unpleasantness invoked by shaking images, we carried out a large amount of psychological estimation experiments. As the result of analysis about the experimental data, we derived out the relationship between physical characteristics of images and the unpleasantness.

抄録全体を表示

In this paper, we propose to extract personalized video features based on the viewer's interest. The viewer's interest toward the video content is modeled in real time by monitoring their pupil size, gazing point and heart rate. The experimental results show that the video summary generated using weighted video features achieves higher precision and recall ratios for video with high narration nature if compared to video summary generated by video features without personalized weight prediction.

抄録全体を表示

Recent developments in video technology enable us to view wider field of view (FOV) images using wide flat panel displays as one would a television monitor. New video systems with high resolution beyond HDTV have been proposed, but it is not clear that the motion picture quality is sufficient if utilizing traditional video system frame rates. In the assessment of picture quality, it is recommended to screen viewers' static visual acuity (SVA), but not dynamic visual acuity (DVA). Furthermore, the methodology to measure the DVA appropriate for the participants in the assessment of picture quality has not yet been established. We experimentally studied the changes in DVA performance by looking at differences in conditions in measuring DVA, as the first step to establish the methodology. We successfully verified the improvement of DVA performances in binocular viewing compared to monocular viewing. We also looked at the differences in DVA performance by the shape of the visual target (Landolt C or E chart).

抄録全体を表示

Border detection between the two monochromatic stimuli (we called them, test and reference) was investigated for a wide range of reference luminance in various surround luminance conditions. Luminance contrast for the border detection shows U shaped curve for each level of surround luminance when plotted against the luminance of reference stimulus. As the surround luminance increases, the U shaped curve shifts toward higher luminance direction, and the lowest contrast value decreases indicating the increase of sensitivity for border detection. Results are discussed in relation to the physiological property of human visual system.

抄録全体を表示

An experimental method to extract optimal fusion parameters of saliency map is proposed. We conducted experiments of singleton search task and measured their eye movements to find a contrast that gives equivalent conspicuity among three features (orientation, intensity, color) from the saccadic latencies. An example of the fusion parameter calculated from the results is also described.

抄録全体を表示

In order to investigate whether different strategies of visual information processing are used by human in intentional learning and impression judgment of faces,trajectories of eye movement were recorded during the two different tasks. Results of the preliminary analysis made on the distribution of eye fixations are introduced.

抄録全体を表示

It is hard to locate every pixel randomly in order to prevent aliasing. The pixels are located randomly toward a specific direction to resolve the problem. The resolving power by the randomly located pixels rises more than two times as compared with that by the use of a low-pass filter.

抄録全体を表示

This paper describes a basic study for adjusting pixel-offset of image sensors using a lens array. The lens array was able to magnify the pixel-offset of image sensors. We experimented with a lens array of a simulation model. As a result, we could easily adjust it with good performance.

抄録全体を表示

We have been investigating an image sensor in which the 2D motion vector of each pixel is detected at very high frame rates. In this paper, we have proposed a new method of on-sensor motion detection and simulated it. Additionally, we fabricated the prototype chip and evaluated it.

抄録全体を表示

We study imaging system for reconstructing high-spatial resolution image using high-speed intermediate images from the image sensor. Developing our proposed system, we designed a test circuit for image sensor. In this paper, first we explain our proposed imaging system, second our test circuit for image sensor.

抄録全体を表示

We successfully developed ultrahigh-speed CCD with maximum frame rate of 2 million frames per second. To estimate maximum frame rate, the equivalent circuit model of the CCD was analyzed. As a result, it was cleared that the driving voltage became blunt with increasing driving speed. The frame rate dependence of the blunt of the driving voltage accorded with the frame rate dependence of the saturation signal level.

抄録全体を表示

Toward an ultraminiature high-performance microphone for next generation, we are studying a silicon electret condenser microphone. Recently we investigated etching and electric characteristics of silicon nitride/oxide electret film, and determined deposition condition of the nitride film. We also developed a fabrication process of electret transducer. A prototype transducer showed sufficiently flat frequency response in audio frequency range.

抄録全体を表示