- 0.25_196_ExtractFrame.sh is used to extract frames from videos;
- genPhi is used to generate sensing matrix;
- generateTrainCNN.m is used to generate image blocks for pre-training key CNN;
- "caffe/model/crX" directory contains all the necessary file for training key CNN. To start the training, run "sh trainCNN_crX.sh";
- GenerateTrainData1ChanVarMix.m is used to generate image blocks for training the whole network;
- extractFeatures_5_25.m is used to extract the intermediate features extracted by the key CNN, and the extracted features are used for training the whole framework;
- "model/crX" directory contains all the files for training the whole framework. The pre-trained key CNN is loaded, further trained with the whole framework using the intermediate feature as input; To start the training the whole framework, run "th trainVideoNet.lua".
cust1105/gpu-video-encode
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
gpu video-encode( 最大提高 Maximum improvement 5dB)
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published