Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/vpanjeta/deep-object-removal
Using cGANs to remove objects from a photo
https://github.com/vpanjeta/deep-object-removal
cgans deep-learning deep-neural-networks deep-object-removal gans neural-network object-removal
Last synced: about 3 hours ago
JSON representation
Using cGANs to remove objects from a photo
- Host: GitHub
- URL: https://github.com/vpanjeta/deep-object-removal
- Owner: VPanjeta
- License: apache-2.0
- Created: 2018-09-27T05:03:05.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2018-10-10T05:19:51.000Z (about 6 years ago)
- Last Synced: 2024-01-08T02:10:59.640Z (10 months ago)
- Topics: cgans, deep-learning, deep-neural-networks, deep-object-removal, gans, neural-network, object-removal
- Language: Python
- Size: 57.9 MB
- Stars: 102
- Watchers: 5
- Forks: 25
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Deep Object Removal
Image completion is a challenging problem because it requires a high-level recognition of scenes. This project tries to achieve object removal from images and get the base image reconstructed based on surrounding colours and objects using conditional GANs.
# Overview
This project is an implementation of cGANs discussed in the paper for [\[General Image Completion\]](https://www.dropbox.com/s/e4l19y9ggqqk2yf/0360.pdf?dl=1)
The models are tweaked a little and implemented to remove objects from images and reconstruct the image without the object.
# Example Usage
## Hot Keys
\[Esc\]: To quit the windowed application.
\[f\]: To filter out the masked object.
\[n\]: To go to the next image.
\[r\]: To refresh and undo all the masking in the current image.
# Description
## Files
### images/
The folder that contains the images to be used in the project. Currently the project requires images of dimensions 400 x 400 which can be changed in the main.py file.
### model/
This folder contains the pretrained model that is trained on mscoco dataset and the model definition file which is written in tensorflow.
### main.py
The main file to run the program. The code runs as an OpenCV windowed application.
### requirements.txt
The requirements file for the project
## Installation
To install the dependencies type
```
sudo pip3 install -r requirements.txt
```
## Run
To run the application type
```{python}
python3 main.py
```
This will run the demo as an OpenCV application
## Dependencies
The project requires the following packages:
OpenCV and OpenCV_python 3.3.0.10
Tensorflow 1.10.1
Numpy 1.13.3