Facial Landmark Detection

Thanks to Kushashwa Ravi Shrimali for inspiring me and mentoring me throughout this project.

Aim & Structure of Repo

• basictechnique/
  • Aim 1: To detect faces for both in images and via using webcam for real time detection.
  • Aim 2: Improve the accuracy of detection model (Cropped the facial region in circular shape.
• data/
  • Haarcascade files for frontal face, mouth, nose and eyes.
• filter1/
  • messyBoxes.py: Blackens the mouth and eyes region and Blurs the background.
  • invisibilityCloak.py: Objects with red color becomes invisible in half of the screen.
  • putStickers.py: Puts glasses and mustache as facial spaces.
  • makeCartoon.py: Cartoonify images.
• landmark_detection/
  • aam/: TODO
  • kazemi/preTrained_image.py && kazemi/viaPreTrainedModel_webCam.py: Detects facial points both for images and via using webcam.
  • lbf/preTrained_image.py && viaPreTrainedModel_webCam.py: Detects facial points both for images and via using webcam.

To Clone the repository, run:

git clone https://github.com/khushi-411/facemark-detection

Dependencies

• opencv
• opencv_contrib