PROJECT TITLE :

Indexing ensembles of exemplar-SVMS with Rejecting taxonomies - 2016

ABSTRACT:

Ensembles of Exemplar-SVMs are used for a large choice of tasks, like object detection, segmentation, label transfer and mid-level feature learning. In order to form this method effective though a giant assortment of classifiers is required, that usually makes the analysis phase prohibitive. To overcome this issue we exploit the joint distribution of exemplar classifier scores to make a taxonomy capable of indexing every Exemplar-SVM and enabling a fast evaluation of the full ensemble. We experiment with the Pascal 2007 benchmark on the task of object detection and on a simple segmentation task, in order to verify the robustness of our indexing knowledge structure close to the standard Ensemble. We additionally introduce a rejection strategy to discard not relevant image patches for a additional efficient access to the info.