Skip to content

Imaging tools CLI for preprocessing datasets before model training.

License

Notifications You must be signed in to change notification settings

3ee-Games/image-tools

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Image Tools CLI 🖼️

PyPI version Downloads Downloads Downloads

Command line interface for pre-processing images for model training.

Features

  • Download all images from a url

  • Convert images to PNG

  • Resize and crop images

  • Chunk large images into smaller squares

  • Remove images with more than one person

Installation

create a virtual environment and imgtools-cli through pip:

python3 -m venv venv

source venv/bin/activate

pip install imgtools-cli

Usage

ℹ️ Help

python -m imgtools_cli -h

⏬ Download all images from a website

-D {url}, {output directory}

python -m imgtools_cli -D https://www.gutenberg.org/cache/epub/67098/pg67098-images.html /Users/ootie/images

✨ Convert images to PNG files

-I {input directory}

python -m imgtools_cli -I /Users/ootie/image_files

✂️ Resize / Crop images

-r {input directory}, {width}, {height}, {crop_focal_point}, {dnn_model_path}

Using crop focal point:

python -m imgtools_cli -r /Users/ootie/images 512 512 True None

Passing in a haar xml to focal crop faces:

python -m imgtools_cli -r /Users/ootie/images 512 512 True /Users/ootie/models/haarcascade_frontalface_default.xml

➗ Chunk large images into squares

Easily take large images and split them into smaller squares for training.

Example: You may want to train on this image but need to split it into smaller squares for training. Chunk Images

Put it through the chunker and you get this:

chunk 1 chunk 2 chunk 3
 Chunked Image Example  Chunked Image Example  Chunked Image Example

-C {dimensions}, {input_directory}, {output_directory}

python -m imgtools_cli -C 512 /Users/ootie/input /Users/ootie/output

🫂 Hassan People Remover

Uses face detection to remove images with more than one person. Helpful for cleaning source images to be used for Stable Diffusion training.

Example: If your input images have more than one person, the image will be deleted:

 Face Detection

Sample images to test with: https://github.com/hassan-sd/people-remover/tree/main/images

-R {input_directory}, {path_to_cascade_xml}

python -m imgtools_cli -R /Users/ootie/image-tools/images/ /Users/ootie/image-tools/examples/haarcascade_frontalface_default.xml

Ported from: https://github.com/hassan-sd/people-remover