IMAGE SEPERATOR USING FACE FACE_RECOGNITION MODULE AND CV2 AND CVLIB
- The images with faces are stored in face folder.
- The images withhout faces are stored in no_face folder.
- We can use CV2 or cvlib or face recogniton module to process image.
- CV2 is faster than face recogniton module.
- To increase accuracy results i made image classification in cv2 that if both face and eye is detetcted then it is valid image.
- Face recogniton is accurate than CV2.
- cvlib is faster and accurate as it uses GPU.
- Dlib(cnn) is very slow but most accurate .
- Dlib(hog) is fast and accurate.
- Run the script and let the script do the magic for you.
| Modules | Speed | Accuracy |
|---|---|---|
| CV2 | Very fast | Medium |
| Face recognition | Slow | High |
| Cvlib | Fast | High |
| Dlib(cnn Face detector) | Very slow | Very high |
| Dlib(hog face detector) | Fast | Medium |
To install dependencies:
pip install -r requirements.txt
How to run script:
python image_sorter.py
Buy me a coffee : click here