Readaptation of a program implementing the R-CNN Mask through a GUI on Python.
The program works as follows:
- The user will first load a video
- Once this is done, he will be able to choose whether or not he can apply the Mask-RCNN on the video.
- Once the choice is made, by pressing the Start button, the Mask R-CNN will be applied to the whole video. The user has the possibility to stop or pause the video generated by the application of the mask.
- Once the video is finished, he can choose to save the result in video format.
NOTES : It is essential to create a folder named "weights", which will contain the COCO model, which has been used as a weight for our Mask R-CNN. The COCO model can be downloaded from : https://github.com/matterport/Mask_RCNN/releases ( after that, download mask_rcnn_coco.h5)
pycocotools installation:
git clone https://github.com/philferriere/cocoapi.git pip install git+https://github.com/philferriere/cocoapi.git#subdirectory=PythonAPI pycocotools requires Visual C++ 2015 Build Tools.
The GUI has been developped in Python 3.6
Credits : matterport - Mask R-CNN implementation Rob--/mask-rcnn-wrapper