DeepLearing

Search Images by Image, a project about Deep Learing application

Based on E-commerce Sites's Goods Search

Phase 1 : Data collection by web crawler clear -- July 10th, 2015
Phase 2 : Feature extraction by caffe clear -- July 16th, 2015
Phase 3 : Feature transformation into hash code clear -- July 21th, 2015
Phase 4 : Multi-Index Hashing clear -- July 24th, 2015
Phase 5 : Multi-label caffe model training clear -- July 31st, 2015
Phase 6 : Modules connected by socket clear -- Aug 4th, 2015
Phase 7 : Optimization(Fast-RCNN, MIH, Weight-based piecewise caffe model training according to different label info concerned) -- Aug 13th, 2015
Project Completed -- Aug 14th, 2015

Result Screenshot

Training project about Deep Learing

Language:C++ 39.9%Language:Python 31.2%Language:Protocol Buffer 27.0%Language:MATLAB 0.9%Language:Shell 0.6%Language:Makefile 0.4%