This university group project aims to develop a neural network architecture that allows the removal of objects from photos added by the user using deep learning techniques.
Kacper Derlatka Github |
Damian Książek Github |
Łukasz Mazurek Github |
Bartłomiej Guz Github |
All you need to do is
- Download weights mask_rcnn_coco.h5 (https://github.com/matterport/Mask_RCNN/releases/download/v2.0/mask_rcnn_coco.h5) and put it in src folder
- Download weights for inpaiting network https://drive.google.com/drive/folders/1CUQa26Pb_AioJjBYTpX3JkPme3M4WcJZ?hl=pl and put it in generative_inpaiting/model_logs/release_places2_256 folder
- Download http://images.cocodataset.org/annotations/annotations_trainval2017.zip and put it in src/annotations/ folder
- Also you need to use pip install -r re quirements.txt to install other requirements
- Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow - https://github.com/matterport/Mask_RCNN - Waleed Abdulla - 2017
- Generative Image Inpainting with Contextual Attention - https://github.com/JiahuiYu/generative_inpainting - Yu, Jiahui and Lin, Zhe and Yang, Jimei and Shen, Xiaohui and Lu, Xin and Huang, Thomas S - 2018
- Free-Form Image Inpainting with Gated Convolution - https://github.com/JiahuiYu/generative_inpainting - Yu, Jiahui and Lin, Zhe and Yang, Jimei and Shen, Xiaohui and Lu, Xin and Huang, Thomas S - 2018