Abstract

We propose a novel approach to annotating weakly labelled data. In contrast to many existing approaches that perform annotation by seeking clusters of self-similar exemplars (minimising intra-class vari- ance), we perform image annotation by selecting exemplars that have never occurred before in the much larger, and strongly annotated, nega- tive training set (maximising inter-class variance). Compared to existing methods, our approach is fast, robust, and obtains state of the art results on two challenging data-sets – voc2007 (all poses), and the msr2 action data-set, where we obtain a 10% increase. Moreover, this use of nega- tive mining complements existing methods, that seek to minimize the intra-class variance, and can be readily integrated with many of them.