Detecting and recognizing objects in images is one of the most challenging tasks in computer vision, as it seeks to detect subtle objects while ignoring massive numbers of negatives. While deep networks have led to ...