Takekazu Kato

WACL: supporting telecommunications using - wearable active camera with laser pointer

We propose a wearable active camera with laser pointer(WACL) as a human interface device for use intelecommunications. The WACL laser pointer is attached tothe active camera-head and it can point a laser spot whileelevating and panning the camera-head. In our system, aremote instructor can observe around the worker, who iswearing the WACL, independently of the workers motion,and can clearly and naturally instruct the worker in tasksby pointing the laser spot at real objects. This paperdescribes the outline of a telecommunication supportsystem using the WACL and a method to stabilize thecamera and laser pointer independently of the wearersmotion. We have implemented and demonstrated anexample of an application using our system.

Appliance Recognition from Electric Current Signals for Information-Energy Integrated Network in Home Environments

We are developing a novel home network system based upon the integration of information and energy. The system aims to analyze user behavior with a power-sensing network and provide various life-support services to manage power and electric appliances according to user behavior and preferences. This paper describes an electric appliance recognition method using power-sensing data measured by CECU (Communication and Energy Care Unit) which is an intelligent outlet with voltage and current sensors to integrate legacy appliances (which are incompatible with a communications network) with the home network. Furthermore, we demonstrate a prototype home energy management system and examples of services based upon appliance recognition.

Energy on demand: Efficient and versatile energy control system for home energy management

We have been proposing the concept of “i-Energy” as a new energy management scheme to realize efficient and versatile control of e-power flows among decentralized energy generation/storage devices and appliances in homes, offices, and neighboring communities. The i-Energy concept is best characterized by a novel energy control method named “Energy on Demand (EoD).” The novelties of EoD rest in (1) the explicit demand-based power supply control, (2) the best-effort power distribution based on appliance priorities, and (3) the ceiling control of power consumption. With EoD, people can attain the guaranteed reduction of energy consumption without damaging their quality of lives. Moreover, when utility companies are allowed to set and modify ceiling values based on contracts with consumers, EoD systems work as smart demand response systems. This paper first describes the protocol for EoD: power request demands named “Quality of Energy (QoEn)” and appliance priority descriptions. Then, the demand mediation algorithm based on appliance priorities for a single power source is introduced. Experiments using real world everyday-life data in a smart apartment room demonstrated the effectiveness of EoD.

Geometric and Timing Calibration for Unsynchronized Cameras Using Trajectories of a Moving Marker

For 3D measurement of moving objects, timing calibration is important to compensate lags in shutter timing between cameras as well as calibrating 3D positions and poses of them. This paper presents a new method for calibrating the geometric relationship and the lags in shutter timing between unsynchronized cameras from trajectories of a moving marker. We call it geometric and timing calibration. Geometric calibration generally requires many point correspondences between images of the calibration objects. It is, however, difficult to obtain stable point correspondences from images. Our method does not require point correspondences. It uses, instead, a pair of trajectories of a freely moving maker. Our method estimates a fundamental matrix and a lag in shutter timing simultaneously by analyzing the gap between the trajectories. In experiments, our method has attained accuracy of mean errors within 0.8 pixels in image coordinates and timing errors within 0.3 milliseconds. Another experiment using a radio-controlled helicopter has shown that it works well for flying objects in the sky

Simple and robust method for detecting the electrical appliances using markers and programmable logic devices

The detection of appliances and their location is important for the efficient management and automatic operation of appliances and their power consumption. In this paper, we describe a simple and robust method of detecting appliances based on their location by using programmable logic devices (PLDs). The results of simulation and implementation are also included.

K-d decision tree: an accelerated and memory efficient nearest neighbor classifier

Most nearest neighbor (NN) classifiers employ NN search algorithms for the acceleration. However, NN classification does not always require the NN search. Based on this idea, we propose a novel algorithm named k-d decision tree (KDDT). Since KDDT uses Voronoi condensed prototypes, it is less memory consuming than naive NN classifiers. We have confirmed that KDDT is much faster than NN search based classifiers through the comparative experiment (from 9 to 369 times faster).

Adaptive storage battery management based on the Energy on Demand protocol

We proposed the concept of i-Energy [1] as a novel smart demand-side energy management scheme to realize efficient and versatile control of e-power flows among decentralized energy generation/storage devices and appliances in homes, offices, and neighboring communities. To embody the concept, we developed a novel energy management method named Energy on Demand (EoD) for a single power source [3]. This paper proposes an adaptive storage battery management method based on EoD. The novelty of the method rests in the demand-based battery capacity design and dynamic charge/discharge control. The effectiveness of the proposed method has been proved with experiments using real life data in the smart apartment room [5].

An Occlusion Robust Likelihood Integration Method for Multi-Camera People Head Tracking

This paper presents a novel method for human head tracking using multiple cameras. Most existing methods estimate 3D target position according to 2D tracking results at different viewpoints. This framework can be easily affected by the inconsistent tracking results on 2D images, which leads 3D tracking failure. For solving this problem, an extension of Condensation using multiple images has been proposed. The method generates many hypotheses on a target (human head) in 3D space and estimates the likelihood of each hypothesis by integrating viewpoint dependent likelihood values of 2D hypotheses projected onto image planes. In theory, viewpoint dependent likelihood values should be integrated by multiplication, however, it is easily affected by occlusions. Thus we investigate this problem and propose a novel likelihood integration method in this paper and implemented a prototype system consisting of six sets of a PC and a camera. We confirmed the robustness against occlusions.

VizWear-Active: distributed Monte Carlo face tracking for wearable active cameras

In this paper we discuss a distributed Monte Carlo (DMC) tracking method which achieves real-time and accurate face tracking for wearable active vision systems. The DMC is an extension of sequential Monte Carlo approaches to a client-server distributed architecture. The client feeds back the results of tracking for use in the control of the wearable active camera with minimal delay, and the server accurately tracks the faces at the same time. In our method, the client is able to complete the tracking processes even when it is unable to communicate with the server Furthermore, more accurate results are obtained when it is able to communicate with the server. We have implemented and demonstrated this method on the VizWear-Active system which includes a wearable active camera.

Cooperative distributed registration for robust face recognition

This paper describes a cooperative distributed registration system for efficient acquisition of a set of facial images required in realization of robust face recognition. In this system, a lot of agents consisting of a camera and a processor are arranged, and the person detection and tracking as well as registration of facial images are integratively realized. By configuring the cooperative mechanism of this system into a three-layer structure, the required communication capacity is reduced and efficient cooperative operation is realized. In the first layer, each agent independently performs the detection and two-dimensional tracking of the target person. In the second layer, the robust three-dimensional tracking is realized by cooperating agents themselves who are tracking a particular target person within an agency. In the third layer, by the cooperation among agencies, the agency configuration is dynamically changed in accordance with the position and registration conditions of the target person, and the effective registration is realized by changing over the target person of each camera. We will present the experimental results based on a system constructed with eight cameras and confirm the effectiveness of the proposed technique.