Search

Shape Context algorithm is a recently proposed good method for object recognition. It is robust under small geometrical distortions and occlusion, invariant under scaling, translation, and change of illumination. However, because of its computational complexity, this algorithm is not suitable for online recognition. In this paper, we proposed an improved Shape Context algorithm which introduces an optimal hierarchical structure and gives new reference points selection criteria. Compared with the original Shape Context algorithm, the improved algorithm keeps the recognition accuracy and reduces the computational cost greatly. The good experimental results reveal that, this new algorithm can be applied to the pedestrian detection system online, which is an important research topic in intelligent transportation.

We have studied spin-dependent transport across magnetic metallic strips deposited on doped Si and GaAs in a wide range of temperatures up to the ambient temperature, and found strong evidence of coherent spin transport through the magnetic metal into the semiconductors at the elevated temperatures. First, spin transport through Si doped GaAs was studied between two cobalt strips separated by a 100 mum. It was found that there was well-defined field modulated resistance variation in low magnetic field range (

This paper describes the design and development of a Cantonese TTVS synthesizer, which can generate highly natural synthetic speech that is precisely time-synchronized with a real-time 3D face rendering. Our Cantonese TTVS synthesizer utilizes a homegrown Cantonese syllable-based concatenative text-to-speech system named CU VOCAL. This paper describes the extension of CU VOCAL to output syllable labels and durations that correspond to the output acoustic wave file. The syllables are decomposed and their initials/finals are mapped to the nearest IPA symbols that correspond to static viseme models. We have authored sixteen static viseme models together with two emotion-based face models. In order to achieve 3D face rendering, we have designed and implemented a blending technique that computes the linear combinations of the static face models to effect smooth transitions in between models. We demonstrate that this design and implementation of a TTVS synthesizer can achieve real-time performance in generation.