A Half-Against-Half (HAH) multi-class SVM is proposed in this paper. Unlike the commonly used One-Against-All (OVA) and One-Against-One (OVO) implementation methods, HAH is built ...
The voice activity detectors (VADs) based on statistical models have shown impressive performances especially when fairly precise statistical models are employed. Moreover, the ac...
- In this paper, we propose a facilitator which finds capable and trustworthy partners on behalf of client users, which helps users form and maintain e-partnerships for electronic ...
Chihiro Ono, Dai Kanemoto, Keesoo Kim, Boyd C. Pau...
This paper tackles the problem of surveillance video content modelling. Given a set of surveillance videos, the aims of our work are twofold: firstly a continuous video is segment...
A central problem in extending the von Neumann architecture to petaflop computers with millions of hardware threads and with a shared memory is defining the memory model [Lam79,...