2018-CVPR-Self-Supervised Learning for Stereo Matching with Self-Improving Ability阅读笔记

2018-CVPR-Self-Supervised Learning for Stereo Matching with Self-Improving Ability阅读笔记_第1张图片

SsSMnet,designed by Yiran Zhong, Yuchao Dai, and Hongdong Li,could be seperated into five part.

Frist part is feature extraction.The structure of it  is genenarl CNN,despite of 

你可能感兴趣的:(Deep,Learning)