FASTER-RCNN (testing): ubuntu14.04+caffe+cuda7.5+cudnn5.1.3+opencv3.0+matlabr2014a Environment Construction Record

Python version of faster-rcnn See my other blog:

PY-FASTER-RCNN (Running the demo): ubuntu14.04+caffe+cuda7.5+cudnn5.1.3+python2.7 Environment Construction record

1. First, you need to configure the environment for compiling Caffe and downgrade GCC to 4.7. See: ubuntu14.04 installation cudnn5.1.3,opencv3.0, compiling caffe and MATLAB and Python interface Process records (sorry , which I wrote myself)

2. Clone Source:

git clone--recursive https://github.com/shaoqingren/faster_rcnn

3. Clone Author's Caffe source code (remember must be the author's, otherwise run MATLAB program error, I think this truth should be very clear). But this step should be done in the recursive clone above.

4. Download the trained model on the Baidu Cloud link provided by the author, of course, "Run to fetch_data/fetch_faster_rcnn_final_model.m download our trained models", but the speed will be very slow.

5. Enter Faster_rcnn/external/caffe, copy the makefile.config before compiling Caffe, also can copy the Makefile.config.example under the current folder, remove. example suffix .

CD External/caffe

6. Modify the Makefile.config file and add the MATLAB path. The important part of my. config file is as follows:

USE_CUDNN: =1opencv_version:=3Cuda_dir:=/usr/local/Cudablas:=Atlas Matlab_dir:=/usr/local/matlab/R2014aanaconda_home:= $ (HOME)/Anaconda2python_include:= $ (anaconda_home)/include $ (anaconda_home)/include/python2.7$ (anaconda_home)/lib/python2.7/site-packages/numpy/core/include Python_lib:= $ (anaconda_home)/Libinclude_dirs:= $ (python_include)/usr/local/Includelibrary_dirs:= $ (python_lib)/usr/local/lib/usr/Lib Use_pkg_config:=1

7 Start compiling caffe and MATLAB interfaces

 Make  Clean  Make-J8 make Matcaffe

8 Run the demo according to the testing steps provided by the author:

1. Runfaster_rcnn_build.m
2. Runstartup.m
3. Run to experiments/script_faster_rcnn_demo.m test A, single demo image.

9 Here it is merit perfection. The approximate results are as follows:

FASTER-RCNN (testing): ubuntu14.04+caffe+cuda7.5+cudnn5.1.3+opencv3.0+matlabr2014a Environment Construction Record