Main characters (persons) extraction into new (desired) background.
This project mainly aims to change the background (Just need an image of your desired background) of input file (Image or video) while keeps the main charactors (person) same as input file. This project uses MASK R-CNN for object detection.
Demo 1 | Demo 2 |
---|---|
![]() |
![]() |
Input Video | Output Video |
---|---|
![]() |
![]() |
numpy
scipy
cython
h5py
Pillow
scikit-image
tensorflow-gpu>=1.3 (Tested on 1.13)
keras
jupyter
matplotlib
imgaug (It requires shapely to be installed)
IPython[all]
opencv-python
Ready the environment by installing all the packges. If you are facing any issues in setting up the environment then you can refer to Mask RCNN page. This project doesn't need any extra packages to run. Before running, please download mask_rcnn_coco.h5 file by clicking here and put it into the main directory
Now, you can run the command "python multi_person.py -h" to know the command line arguments requirements.
Mask RCNN