Scanner: efficient video analysis at scale

Information

• Paper topic: Images
• Software type: Code
• Able to run a replicability test: True
• Replicability score: 4
• Software language: C/C++, Python
• License: Apache License 2.0
• Build mechanism: Makefile, CMakeLists, Not applicable (python, Matlab..)
• Dependencies: tensorflow/caffe/opencv/eigen3/ffmpeg/boost/openpose/halide/storehouse/hwang/pybind/grpc/protobuf/libpqxx/cuda/cudnn
• Documentation score {0,1,2}: 1
• Reviewer: Julie Digne <julie.digne@liris.cnrs.fr>
• Time spent for the test (build->first run, timeout at 100min): 50min

Source code information

• Code URL: https://github.com/scanner-research/scanner
• Git commit hash: 645fb2388341bc90300e2645fe4f9d0374baa537
• Software Heritage link:

Comments

I used the docker setting, it worked well (cpu). Was not able to install from source because of a problem with hwang (installed by deps.sh although optional dependency). A jupyter notebook extremely detailed is provided and helps understand the code, so it would be very easy to adapt. I did not reproduce all the results in the paper (grayscale conversion worked well with docker).

If you want to contribute with another review, please follow these instructions.

Please consider to cut/paste/edit the raw JSON data attached to this paper.