I've tried using the trig ratios to rotate the face based on a single eye detected with the Viola-Jones algorithm. Unfortunately, the Viola-Jones algorithm has not been giving me reliable eye detections, so, it's affecting my trig calculations. I've also tried extracting SURF features/keypoints from a frontal image and using it as a baseline for aligning other images, but it's still not working. I want it automated, I have a large number of images (over 20,000) and I can't image aligning them manually, more so, I won't manually align test images in the real application.
I await your contributions
Thank you.
Refer article "DeepFace: Closing the Gap to Human-Level Performance in Face Verification" by Yaniv Taigman. The authors have used LFW face database.
I recommend you to take a look to this feature which works very well:
http://docs.opencv.org/trunk/d7/d8b/tutorial_py_face_detection.html
- Badges
- Science topic
- Similar topics
- Computer Science
- Images