image2layout-computer-vision 0.0.1

image processing and stuff

image2layout-computer-vision

Computer Vision related modules for image2layout

Installation

1. Build and run Docker container, will contain the necessary requirements for all modules

sudo docker build --tag cv -f Dockerfile .

sudo docker run -it -v $(pwd):/app cv bash

2. OR, Install Conda

curl https://repo.anaconda.com/archive/Anaconda3-2023.03-1-Linux-x86_64.sh -o ~/conda.sh
bash ~/conda.sh -b -f -p /opt/conda
rm ~/conda.sh
conda init --all --dry-run --verbose

OCR - Text Detection

Installation [CPU]

Follow instructions in README to build and run docker container for all modules

Or install using conda locally:

Python [conda] + tesseract

sudo apt install tesseract-ocr libtesseract-dev -y

conda create -n ocr python=3.8 -y
conda activate ocr

conda install -n ocr pytorch=1.10 torchvision -c pytorch -y

python -m pip install Pillow pandas numpy
python -m pip install detectron2 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.10/index.html
python -m pip install pyyaml==5.1 chardet pytesseract
python -m pip install --upgrade datasets transformers

Usage

from main import detect_text

# returns an ImageBoxes object
imageboxes = detect_text('path/to/image.png')

# draw boxes
imageboxes.draw_anno()

# dataframe
imageboxes.df

# boxes (x0, y0, x1, y1)
boxes = np.array(list(imageboxes.df['box'])).astype(int)

Color Extractor

Installation

Python

python -m pip install Pillow pandas numpy scikit-learn

Usage

from PIL import Image
from main import ColorExtractor

image = Image.open('path/to/image.png').convert('RGB')
color_extractor = ColorExtractor(image)
color_extractor.colors
color_extractor.color_bg    # background color
color_extractor.color_fg    # foreground color