Towards the goal of realizing a generic automatichuman activity recognition system, a new formalism is proposed. Activities are described by a chained hierarchical representation ...
There are many computer vision algorithms developed for visual (scene and object) recognition. Some systems focus on involved learning algorithms, some leverage millions of trainin...
The problem of recognizing actions in realistic videos
is challenging yet absorbing owing to its great potentials
in many practical applications. Most previous research is
limit...
Jintao Li, Ju Sun, Loong Fah Cheong, Shuicheng Yan...
When exposed to environmental noise, speakers adjust their speech production to maintain intelligible communication. This phenomenon, called Lombard effect (LE), is known to consi...
In this paper, we present a scheme towards recognition of English character in multi-scale and multi-oriented environments. Graphical document such as map consists of text lines w...