We propose to bridge the gap between Random Field (RF) formulations for joint categorization and segmentation (JCaS), which model local interactions among pixels and superpixels, ...
Interactive programs, such as GUIs or spreadsheets, often maintain dependency information over dynamically-created networks of objects. That is, each imperative object tracks not ...
Neel R. Krishnaswami, Lars Birkedal, Jonathan Aldr...
Visualization is often invaluable to understand the behavior of optimization algorithms, identify their bottlenecks or pathological behaviors, and suggest remedial techniques. Yet ...
—We present in this paper an integrated solution to rapidly recognizing dynamic objects in surveillance videos by exploring various contextual information. This solution consists...
Xiaobai Liu, Liang Lin, Shuicheng Yan, Hai Jin, We...
Most of the Web-based methods for lexicon augmenting consist in capturing global semantic features of the targeted domain in order to collect relevant documents from the Web. We s...