and The Role of Hierarchical Group Sparsity in Semantic Learning of Images

As computer vision algorithms have become more and more sophisticated in the task of object detection, semantic learning remains a harder goal to achieve. Let us assume that a good object detection/feature extraction algorithm is given, how does one build on that information to gain context about the semantic label of the scene? This talk will first briefly go through current methods used in solving semantic learning of images. I will then talk about our approach and present some early results.