Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/samjuk/facial-grouping
Python facial extraction and grouping
https://github.com/samjuk/facial-grouping
age-detection face-detection face-recognition gender-detection hacktoberfest hacktoberfest2020 opencv
Last synced: 6 days ago
JSON representation
Python facial extraction and grouping
- Host: GitHub
- URL: https://github.com/samjuk/facial-grouping
- Owner: SamJUK
- Created: 2020-05-17T00:57:44.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2022-11-09T08:09:48.000Z (over 2 years ago)
- Last Synced: 2023-02-26T23:36:12.636Z (almost 2 years ago)
- Topics: age-detection, face-detection, face-recognition, gender-detection, hacktoberfest, hacktoberfest2020, opencv
- Language: Python
- Homepage:
- Size: 84 MB
- Stars: 1
- Watchers: 2
- Forks: 2
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Python Face Extraction
This project is made up of two components, a video face extractor and an age & gender detector. The idea of the project is to be able to run community made videos through it and classify by the age of the people within them. So those of certain age groups can be flagged for manual review.
```sh
$ python face_detector.py --video=videos/face-demographics-walking.mp4 --tolerance=.55 --blur=100
$ python age_gender_detect.py --images=faces/videos_face_demographics_walking_mp4/E1
```- If classifying same person as different people, raise tolerance
- If classifying different people as the same, lower tolerance## Arguments
- **video:** Path to video file
- **frames:** What frame multiplier to process (e.g 0.5 = Half the frames, 0.1 = tenth of the frames)
- `0.01` = Default
- `-1` = Every frame
- `High Number` = Fast, Less Accurate
- `Low number` = Slower, more accurate
- **matchaccuracy:** What percentage of faces to match in a person group to be classified the same
- `0.15` = Default
- **blur:** Integer of blur intensity (only affects preview)
- **framesize:** float from `0 - 1` that scales the image size (smaller is faster but less accurate)
- **tolerance:** Face detection match tolerance## Networks
The Age & Gender detection models are based from the [AgeGenderDeepLearning](https://github.com/GilLevi/AgeGenderDeepLearning) project.