Additive Kernels and Explicit Embeddings for Large Scale Computer Vision Problems

Organizers: Jianxin Wu (Nanyang Technological University, Singapore), Andrea Vedaldi (Univ. of Oxford, UK), Subhransu Maji (TTI Chicago, USA), Florent Perronnin (Xerox Research Center Europe)
Duration: half day
Abstract: It is generally accepted in our community that: in many vision tasks, more training images will usually lead to better performance. Furthermore, recent advances have shown that additive kernel and explicit embeddings are the best performers in most visual classification tasks–a fact that has been repeatedly verified by various papers and research-oriented public contests (e.g., the ImageNet Large Scale Visual Recognition Challenge.) In this tutorial, we will introduce the theories, applications, algorithms, software, and practical issues of using additive kernels and explicit embeddings in various computer vision domains, especially when the problem scale is very large.