Kazunori Sugiura

Design and implementation of DV stream over Internet

An Internet Protocol (IP) based high-quality real-time audio-visual (AV) communication tool has been developed. A digital video (DV) camcorder and VCR, being consumer market products, are adopted as the AV codec. The AV devices, which have an IEEE-1394 (Firewire) interface, are connected to the IP host with its interface. The original packet format is adopted by simply encapsulating the isochronous stream packet of the DV with UDP/IP. In the highest quality of communication, the system consumes over 35 Mbps of network bandwidth. The system can also change the bandwidth adaptively for end-to-end network conditions. The DV communication effort was performed using the APAN TransPAC link between the USA and Asia (Japan and Korea).

Design and implementation of distributed and scalable multimedia signage system

Digital signage systems have found many interesting applications in the areas of advertising, entertainment and education. There are a variety of digital signage architectures in use. Through this paper we present Infoshare, a distributed and scalable multimedia signage system. Infoshare is based on the architecture of a collection of autonomous signage players connected with servers via a web service interface. The signage players display XAML based screens with rich signage contents which are scheduled based on location and time. This paper provides an overview of the Infoshare system that is deployed as the news signage system of Graduate School of Media Design, Keio University.

AnyControl -- IoT Based Home Appliances Monitoring and Controlling

Technologies for Internet of Things (IoT) such as sensor, network and data processing are advancing rapidly. Thanks to the evolution, many potential applications have been developed in the fields of analysis and automation. In this paper, a platform was designed to connect sensor data with users daily life. As an application of it, AnyControl - a home appliances monitoring and controlling system was implemented. Although home appliances are becoming more intelligent day by day. Not only are manufacturers promoting new smart appliances, there are also many smartphone oriented remote controller products. However current products always have platform compatibility problems, additionally, user interaction in such systems are becoming more and more complex. This work proposes an approach to enhance old appliances and the controlling experience through an IoT based Home Appliance Controlling System. With sensors the appliances can be controlled environment condition trigger. The sensor data are processed by single-board computer and delivered to mobile applications through wireless connection. The results of implementation and experimentation has shown the proposed system and platform can provide more IoT application possibilities daily life.

GazeSim: simulating foveated rendering using depth in eye gaze for VR

We present a novel technique of implementing customized hardware that uses eye gaze focus depth as an input modality for virtual reality applications. By utilizing eye tracking technology, our system can detect the point in depth the viewer focusses on, and therefore promises more natural responses of the eye to stimuli, which will help overcoming VR sickness and nausea. The obtained information for the depth focus of the eye allows the utilization of foveated rendering to keep the computing workload low and create a more natural image that is clear in the focused field, but blurred outside that field.

In the Eye of the Beholder: The Impact of Frame Rate on Human Eye Blink

We introduce a study investigating the impact of high frame rate videos on viewers eye blink frequency. A series of videos with varying combinations of motion complexities and frame rates were shown to participants, while their eye blinks were counted with J!NS MEME (smart eye wear). Lower frame rates and lower motion complexity caused higher blink frequencies, which are markers for stress and emotional arousal.

Adaptive Rate Control with Dynamic FEC for Real-Time DV Streaming

For higher streaming quality, a data sender adjusts the data transmission rate according to the network condition between the sender and receiver. The sender and the receiver exchange information about the network condition to decide the appropriate transmission rate. However, monitoring each data flow in real time is difficult, and controlling the sender to adjust the best quality for each receiver in a heterogeneous environment is a real challenge. In this paper, we study packet loss patterns and FEC recovery rate upon data transmission in a congested network, and define an adaptive rate control mechanism that dynamically adjusts the transmission rate and the FEC encoding rate. We then show our adaptive DV Transmission System (DVTS) that supports appropriate rate control for provisioning multimedia streaming with the best possible quality, and evaluate the system on top of our testbed network.

TelescopeVisualizer: a real-time internet information visualizer with a flexible user interface

Visualizing the Internet activities has been an interesting and challenging research issue. Many successful efforts were concentrated around visualizing a single aspect of the Internet. However such fixed application user interfaces cannot cope with visualizing the diversity of network information and ever changing nature of the Internet. We present TelescopeVisualizer, a scalable real-time Internet information visualizer that allows for querying and visualizing multiple aspects of the Internet interactions. TelescopeVisualizers XAML based flexible user interface avoids tight coupling of visualization application with the rest of the system making it possible to visualize Internet information in realtime. It allows users to upload custom query layouts, post queries and change the visualizations as they require. Using a query traffic generator application, we demonstrate scalability of TelescopeVisualizer system in different hardware platforms.

On the delay and quality of DV transmission systems using ATM networks

The transmission of DV data which needs 28 Mbps traffic has been possible, because public ATM networks, like Japan Giga-bit Network (JGN) can be used. The DV system not only provides good audio and video quality, but also has a small processing delay. The DV system is suitable for interactive systems like teleconferencing for the small delay. We report the result of measurements of the delay time of DV systems on JGN. DVTS over IP and Sony SEU-TL100 over ATM are measured. The delay time of the systems are 90-220 msec including the NTSC/DV converter. The delay time of the ATM network is 11 msec for Tokyo-Osaka where the distance and routes were varied. The jitters was also varied for congestion analysis. The jitter make the quality of DV decrease rapidly when system buffers are small.

Facial Thermography for Attention Tracking on Smart Eyewear: An Initial Study

We are describing the first step towards the development of an unobtrusive open eyewear system for attention tracking in daily life situations. We are logging thermographic data from infrared imaging and electrooculographic readings from off-the-shelf smart glasses and measure cognitive engagement of people in different situations. We are identifying new potential areas on the face for contactless IR temperature sensing. Attached to smart glasses and in combination with the EOG potentials we can monitor the wearers facial temperature changes, eye movement and eye blink in everyday situations, which is a major step towards becoming able to measure attention in unconstrained settings, and thus make it manageable.

Eye blink as an input modality for a responsive adaptable video system

We propose a unique system that allows real-time adaption of video settings to a viewers physical state. A custom made program toggles between videos according to the average eye blink frequency of each viewer. The physical data is harnessed with J!NS MEME smart glasses that utilize electrooculography (EOG). To the best of our belief, this is the first adaptable multimedia system that responds in real time to physical data and alters technical settings of video contents.