I am looking for already trained haarcascades of facial fiducial points (left corner of left eye, right corner of left eye, left corner of right eye, right corner of right eye, left corner of mouth, right corner of mouth, left, center and right side of nose).
Does anyone know where can I download an already trained haarcascades to use with OpenCV's VJ function?
相关问题
- How to get the background from multiple images by
- Try to load image with Highgui.imread (OpenCV + An
- CV2 Image Error: error: (-215:Assertion failed) !s
- Is it a bug of opencv RotatedRect?
- How do I apply a perspective transform with more t
相关文章
- opencv fails to build with ipp support enabled
- Code completion is not working for OpenCV and Pyth
- Face unlock code in Android open source project?
- How to compile a c++ application using static open
- Why cv2.rectangle sometimes return np.ndarray, whi
- Fastest way to compute image dataset channel wise
- TypeError: unsupported operand type(s) for +: '
- Error installing OpenCV on Ubuntu 16.04
If I m not so badly mistaken, HaarCascade works with objects not with specific points. There are already classifiers for nose,eye and mouth within opencv data folder. You can detect eyes,nose and mouth with them, than you can find corners of them by using specifically well and wise crafted detection algorithms (use your imagination and mind).
Here an example;
This example finds the eyes in the loaded image. Finding mouth and nose is also similar. It finds them as rectangle and we draw these rectangles to image. I don't know if these rectangles are enough for you but if you want more specific points, like center of nose; you will need to do more processing.
In this Answer I found two links to libraries to find facial feature points based on the opencv:
Here are the two libraries: