You need to clone this repo first and create conda virtual environment.
git clone https://github.com/nguyen-brat/text-remove.git
cd text-remove
conda create -n inpaint python==3.8
conda activate inpaint
sudo apt-get install libjpeg-dev zlib1g-dev
sudo apt install unzip
pip install -r requirements.txt
cd craft_pytorch
wget --no-check-certificate 'https://docs.google.com/uc?export=download&id=1Jk4eGD7crsqCCg9C9VjCLkMN3ze8kutZ' -O "craft_mlt_25k.pth"
gdown 1XSaFwBkOaFOdtk4Ane3DFyJGPRw6v5bO
cd ..
cd lama
curl -LJO https://huggingface.co/smartywu/big-lama/resolve/main/big-lama.zip
unzip big-lama.zip
cd ..
bash config/text_detection.sh -s path/to/the/folder/remove/text \
-t path/to/the/folder/saved/result
bash config/text_detection.sh -s test_folder -t target_test
streamlit run app.py
- config/: Configuration files for the project.
- craft_pytorch/: Contains code related to CRAFT text detector using PyTorch.
- image/: Directory for storing image files used in the project.
- lama/: Directory related to the LaMa inpainting model.
- mmocr/: Contains files related to MMOCR (Multimodal OCR) toolkit.
- myenv/: Virtual environment for the project dependencies.
- notebook/: Jupyter notebooks for experimentation and prototyping.
- target_test/: Directory for target test files.
- test_folder/: Directory for test files.
- .gitignore: Git ignore file to specify which files and directories should be ignored by Git.
- app.py: Main application script.
- draw_bb.py: Script for drawing bounding boxes.
- README.md: The file you are currently reading.
- requirements.txt: List of project dependencies.