Object Detection《SSP-Net》笔记（2）

Spatial pyramid pooling in deep convolutional networks for visual recognition 说明：compute the feature maps from the entire image only once, and then pool features in arbitrary regions (sub-images) to g

>>阅读原文<<