favorite4 Santosh Kumar Divvala, Derek Hoiem, James Hays, Alexei A.
favorite121There are considerable works in categorical object detection [6, 8, 15, 33, 35] which study the impact of different object-level properties (such as occlusion, aspect ratio, viewpoint change) to the performance of categorical object detectors.
favorite9, many object proposal methods have been proposed [2, 3, 4, 9, 18, 20, 23, 26, 27, 32, 37] and tested on various large scale datasets [11, 22, 29], and their overall detection rates versus different thresholds or window number have also been reported.
favorite0Our study reveals the limitations of existing methods in terms of non-iconic view, small object size, low color contrast, shape regularity etc.
favorite6Abstract Object proposal has become a popular paradigm to replace exhaustive sliding window search in current top-performing methods in PASCAL VOC and ImageNet. Specifically, we examine the effects of object size, aspect ratio, iconic view, color contrast, shape regularity and texture.