Zzzbang / caffe-sal-ensemble

Make one caffe, test all methods. This caffe is an ensemble version of several salient object detection repos.

caffe-sal-ensemble

Introduction

This repository is an ensemble version of several salient object detection repos.

• AFNet
• Amulet
• C2SNet
• DHSNet
• DSS
• PiCANet
• RAS
• SRM
• UCF

Installation

The code is built on Ubuntu 14.04 with CUDA 8.0 and OpenCV 3.1.

Usage

1. Clone the repository:

   git clone https://github.com/ArcherFMY/caffe-sal-ensemble.git

2. Build Caffe and matcaffe:

   cd $CAFFE_SAL_ENSEMBLE_ROOT && cd caffe
   cp Makefile.config.example Makefile.config 
   vim Makefile.config #edit Makefile.config for your environment
   make all -j
   make matcaffe

3. Run the test demo:

   • The trained model is put in folder pretrained_models: -> xxx.caffemodel (download their released pretrained models and change the name to match their method name.)

   • The network model is put in folder prototxts: -> deploy-xxx.prototxt

   • The test images can be found in test-Image

   • Then run test_xxx.m using matlab, you can get the saliency maps in folder results/xxx/

Performance

You could use this evaluation tool box.

License

This code is released under the MIT License (refer to the LICENSE file for details).

Contact

Please contact [e-mail]