Skip to content

Report generator for eyeMon's eye blink detection methods

License

Notifications You must be signed in to change notification settings

predkambrij/eyeMonTester

Repository files navigation

eyeMonTester

Testing framework for https://github.com/predkambrij/eyeMon

EyeMonTester runs eyeMon's methods against recorded videos. EyeMonTester can generate overall report (in CSV format) and draw debugging graphs from intermediate results.

Prepare Docker images:

  • build eyeMon Docker image (see the instructions in its README.md
  • git clone [email protected]:predkambrij/eyeMonTester.git
  • docker build --build-arg ARG_UID=$(id -u) --build-arg ARG_GID=$(id -g) --build-arg ARG_DOCKERGID=$(cat /etc/group | grep docker | awk -F\: '{print $3}') -t predkambrij/eyemontester eyeMonTester/

Prepare testing videos:

  • Add videos (optionally with annotations) in this directory https://github.com/predkambrij/eyeMon/tree/master/posnetki
    • You can get example videos (with annotations) here:
    • wget http://www2.fiit.stuba.sk/~fogelton/acvr2014/eyeblink8.zip -o /tmp/eyeblink8.zip
    • mkdir -p eyeMon/posnetki/sk/eyeblink8
    • echo "4dd946a9318b0f1062e95cf8b658a456 /tmp/eyeblink8.zip"|md5sum -c -
    • unzip /tmp/eyeblink8.zip -d eyeMon/posnetki/sk/eyeblink8/
  • Update videos and videoRange variables (in prepareVideosList function) so that it refers to your video locations
    • example:
    • videos = eyeblink8HandInTheFirstFrames+eyeblink8 # videos that are available to process
    • videoRange = range(len(videos)) # videos you would like to process (you can select subset of available videos)

Start the container:

  • docker run -it --rm -e DIRPREF=$(pwd) -v $(pwd)/eyeMon/_1OpenCVopticalflow/src/main/jni/main_settings_testpy.cpp:/tmp/main_settings_testpy.cpp -v $(pwd)/eyeMon/_1OpenCVopticalflow/src/main/jni/common_settings_testpy.cpp:/tmp/common_settings_testpy.cpp -v $(pwd)/eyeMon/posnetki:/posnetki -v $(pwd)/eyeMonTester:/eyeMonTester -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY -v /var/run/docker.sock:/var/run/docker.sock predkambrij/eyemontester bash

Generating intermediate results:

  • cd /eyeMonTester
  • python main.py

See the results:

  • python misc.py -r # generate CSV report
  • python misc.py -g # Show debugging graphs
  • python misc.py -h # Show help (other options)

About

Report generator for eyeMon's eye blink detection methods

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages