We introduce a new image representation that encompasses both the general layout of groups of quantized local invariant descriptors as well as their relative frequency. A graph of...
People re-detection aims at performing re-identification of people who leave the scene and reappear after some time. This is an important problem especially in video surveillance ...
—Traditional connectionist classification models place an emphasis on learned synaptic weights. Based on neurobiological evidence, a new approach is developed and experimentally ...
This paper presents a parse-and-paraphrase paradigm to assess the degrees of sentiment for product reviews. Sentiment identification has been well studied; however, most previous ...
Decision trees have been widely used for online learning classification. Many approaches usually need large data stream to finish decision trees induction, as show notable limitat...