https://github.com/jonashoechst/facecropper
facecropper: detect faces, grayscale and create circular images
https://github.com/jonashoechst/facecropper
circle circular crop crop-image detection face facecropper head image processing
Last synced: 8 months ago
JSON representation
facecropper: detect faces, grayscale and create circular images
- Host: GitHub
- URL: https://github.com/jonashoechst/facecropper
- Owner: jonashoechst
- License: gpl-3.0
- Created: 2020-04-30T11:44:11.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2024-10-08T11:33:53.000Z (over 1 year ago)
- Last Synced: 2025-04-23T16:09:26.193Z (about 1 year ago)
- Topics: circle, circular, crop, crop-image, detection, face, facecropper, head, image, processing
- Language: Python
- Homepage:
- Size: 2.25 MB
- Stars: 3
- Watchers: 1
- Forks: 1
- Open Issues: 1
-
Metadata Files:
- Readme: Readme.md
- License: LICENSE
Awesome Lists containing this project
README
facecropper
===
`facecropper` is tool to extract faces from images. It facilitates OpenCV to detect faces, grayscale and create circular images.
There are multiple settings, such as the padding, background color, or grayscaling:
```bash
facecropper -v -o docs/lenna_0.png docs/lenna.png
facecropper -v -o docs/lenna_1.png docs/lenna.png --grayscale
facecropper -v -o docs/lenna_2.png docs/lenna.png --grayscale --padding 1.0
facecropper -v -o docs/lenna_3.png docs/lenna.png --padding 1.0 --color "(122, 90, 50, 255)"
```
## Installation
`facecropper` comes a python package and thus can be installed using pip.
```bash
git clone https://github.com/jonashoechst/facecropper.git
pushd facecropper
pip install .
```
## Usage
```bash
$ facecropper -h
usage: facecropper [-h] [--cascade CASCADE] [-o OUTPUT] [-p PADDING] [-s SIZE]
[-g] [-v] [-c COLOR]
image [image ...]
Detect and crop faces from an image.
positional arguments:
image path to input image
optional arguments:
-h, --help show this help message and exit
--cascade CASCADE face detection cascade to be used by OpenCV
-o OUTPUT, --output OUTPUT
Output path template, evaluates placehoders:
{path} -> original file path,
{name} -> original file name,
{ext} -> original file extension,
{i} -> index of detected face
-p PADDING, --padding PADDING
relative space around recognized face (> 0), default=0.3
-s SIZE, --size SIZE export image resolution height / width, default=200
-g, --grayscale grayscale cropped image
-v, --verbose increase verbosity (may be applied multiple times)
-c COLOR, --color COLOR
background color for circular cutout, BRG(A)-format, default: (255, 255, 255, 0)
```
### Cascades
The cascades used by OpenCV can be
- user-supplied by giving a path to the respective xml file or
- be selected from the project:
```
haarcascade_eye.xml
haarcascade_eye_tree_eyeglasses.xml
haarcascade_frontalcatface.xml
haarcascade_frontalcatface_extended.xml
haarcascade_frontalface_alt.xml
haarcascade_frontalface_alt2.xml
haarcascade_frontalface_alt_tree.xml
haarcascade_frontalface_default.xml
haarcascade_fullbody.xml
haarcascade_lefteye_2splits.xml
haarcascade_licence_plate_rus_16stages.xml
haarcascade_lowerbody.xml
haarcascade_profileface.xml
haarcascade_righteye_2splits.xml
haarcascade_russian_plate_number.xml
haarcascade_smile.xml
haarcascade_upperbody.xml
```
## Example
```bash
$ facecropper -v -o docs/lenna_3.png docs/lenna.png --padding 1.0 --color "(122, 90, 50, 255)"
INFO: Loading /Users/hoechst/Projects/facecropper/facecropper/haarcascades/haarcascade_frontalface_default.xml
INFO: Processing docs/lenna.png, resolution: 512x512
INFO: detected face of size 173x173
INFO: Exporting docs/lenna_3.png, grayscale: False
```