How good are detection proposals, really?
This work addresses the problem of understanding detection proposal trade-offs for researchers and practitioners in computer vision, but it is incremental as it focuses on analysis rather than introducing new methods.
The paper analyzed ten object proposal methods and four baselines to evaluate their trade-offs in object detection, finding common weaknesses and providing guidance for method selection.
Current top performing Pascal VOC object detectors employ detection proposals to guide the search for objects thereby avoiding exhaustive sliding window search across images. Despite the popularity of detection proposals, it is unclear which trade-offs are made when using them during object detection. We provide an in depth analysis of ten object proposal methods along with four baselines regarding ground truth annotation recall (on Pascal VOC 2007 and ImageNet 2013), repeatability, and impact on DPM detector performance. Our findings show common weaknesses of existing methods, and provide insights to choose the most adequate method for different settings.