PyQt-Image-Caption

This codebase contains a pyqt application to create image captioning dataset to train image captioning models.

Docker Image

build image

docker build --no-cache -t desalef/pyqt-caption-dataset .

run container

xhost +
docker run --rm -it \
	-e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix \
  --ipc=host --volume="$HOME/.Xauthority:/root/.Xauthority:rw" \
  --userns=host \
  --net=host \
  desalef/pyqt-caption-dataset:latest /bin/bash

run main.py inside the container

python3 src/main.py

Note:

To caption images inside the folder you need to mount the folder into the docker container when you run the container. simply add " -v path/to/images:/usr/app \ " after " --net=host \ " on run command provided above.

TODO:

• export captions into different dataset format
• Undo and Redo changes
• Enter for save and load next image

Sample Screenshot

Dirctory Structure

├── assets
│   ├── arrow-left.png
│   ├── arrow-right.png
│   └── ...
├── docker
│   ├── build.sh
│   └── run-docker.sh
├── src
│   ├── __init__.py
│   ├── actions.py
│   └── main.py
├── utils
│   ├── __init__.py
│   └── utils.py
├── Dockerfile
├── __init__.py
├── LICENSE
├── main.ui
└── README.md