Large Scale Hard Sample Mining with Monte Carlo Tree Search
Large Scale Hard Sample Mining with Monte Carlo Tree Search
复制标题
DOI：
10.1109/cvpr.2016.554
复制
发表时间：
2016-06
期刊：
影响因子：
--
通讯作者：
Olivier Canévet;F. Fleuret
中科院分区：
文献类型：
--
作者：
Olivier Canévet;F. Fleuret
我们研究了一种有效的策略，从非常大的训练集的背景下，对象检测收集假阳性。我们的方法通过使用反映检测器响应的统计规律性的图像集合的分层分解来扩展标准的自举过程。基于该分解，我们的程序使用蒙特卡罗树搜索优先考虑采样到子家庭的图像已被观察到是丰富的误报，同时保持一小部分的采样对未开发的子家庭的图像。由此产生的程序大大增加了假阳性样本的比例相比，一个天真的均匀采样访问的。我们实验应用这个新的过程中收集的100，000张背景图像和行人检测与32，000张图像的人脸检测。我们表明，对于两个标准的检测器，所提出的策略削减了一半的图像访问的数量，以获得相同数量的误报和相同的最终性能。
We investigate an efficient strategy to collect false positives from very large training sets in the context of object detection. Our approach scales up the standard bootstrapping procedure by using a hierarchical decomposition of an image collection which reflects the statistical regularity of the detector's responses. Based on that decomposition, our procedure uses a Monte Carlo Tree Search to prioritize the sampling toward sub-families of images which have been observed to be rich in false positives, while maintaining a fraction of the sampling toward unexplored sub-families of images. The resulting procedure increases substantially the proportion of false positive samples among the visited ones compared to a naive uniform sampling. We apply experimentally this new procedure to face detection with a collection of 100,000 background images and to pedestrian detection with 32,000 images. We show that for two standard detectors, the proposed strategy cuts the number of images to visit by half to obtain the same amount of false positives and the same final performance.