Analyzing videos of human activities involves not only recognizing actions (typically based on their appearances), but also determining the story/plot of the video. The storyline ...
Many modern computer vision systems are built by chaining together standard vision procedures, often in graphical programming environments such as Khoros, CVIPtools or IUE. Typical...
This paper presents a parallel phoneme recognition algorithm based on the continuous Hidden Markov Model (HMM). The parallel phoneme recognition algorithm distributes 3-state HMMs...
To utilize CT or MRI images for computer aided diagnosis applications, robust features that represent 3-D image data need to be constructed and subsequently used by a classificati...
In this paper, we present a novel graph construction method and demonstrate its usage in a broad range of applications starting from a relatively simple single-surface segmentation...