it is based on the https://github.com/matterport/Mask_RCNN.
Therefore, it is necessary to install the pycocotools and the mask-rcnn-coco.h5 model.
I modify the demo.py so that I can save mask image instead of showing in the figure.
It is worth to mention that the specific segmenttation object we catch is the person who might be the main character in the image.
content：
mask：
