Kazuyuki Imagawa

Color-based hands tracking system for sign language recognition

The paper describes a real-time system which tracks the uncovered/unmarked hands of a person performing sign language. It extracts the face and hand regions using their skin colors, computes blobs and then tracks the location of each hand using a Kalman filter. The system has been tested for hand tracking using actual sign-language motion by native signers. The experimental results indicate that the system is capable of tracking hands even while they are overlapping the face.

Real-time face detection with MPEG-4 codec LSI for a mobile multimedia terminal

We propose an intelligent easy-operating human-machine interface incorporating face detection technology for a mobile multimedia terminal. We have developed a face detection algorithm which works on the resources of a mobile phone. This algorithm is robust in outside fields, and simultaneously works with MPEG-4 Simple Profile Level 1 encoding and decoding on an MPEG-4 codec LSI for a mobile videophone. This enables more intelligent interfaces in visual communication such as IMT-2000 videophone services.

Real-Time Tracking of Human Hands from a Sign-Language Image Sequence

We have developed a real-time system which tracks the hands of a person doing sign language. The system enables us to track hands without markers or colored gloves even if the hands overlap the face. First, the system extracts the hand and face regions from the sign-language image sequence using an improved histogram backprojection. Next, the system tracks hands from blobs which are computed from both the extracted image and the time differential image. The system has been tested for hand tracking using both primitive motions and the actual motions of sign-language used by native signers. The experimental results indicate that the system is able to track hands while the hand overlaps the face.

Human Interface. Recognition of Local Features for Camera-based Sign-Language Recognition System.

A sign-language recognition system should use information from both global features, such as hand movement and location, and local features, such as hand shape and orientation. We designed a system that first selects possible words by using the detected global features, then narrows the choices down to one by using the detected local features.In this paper, we describe an adequate local feature recognizer for a sign-language recognition system. Our basic approach is to represent the hand images extracted from sign-language images as symbols corresponding to clusters by using a clustering technique. The clusters are created from a training set of extracted hand images so that images with a similar appearance can be classified into the same cluster in an eigenspace. Experimental results showed that our system can recognize a signed word even in two-handed and hand-to-hand contact cases.

Profile face detection using block difference feature for automatic image annotation

We propose a new profile face detection system for automatic image annotation. For personal images, profile face detection is required because faces turn in different directions. To reduce the numerous false detections seen in previous approaches, we propose a block difference feature. The experiments we describe show the successful outcome of this strategy.