问题
How do people usually extract the shape of the lips once the mouth region is found (in my case using haar cascade)? I tried color segmentation and edge/corner detection but they're very inaccurate for me. I need to find the two corners and the very upper and lower lip at the center. I've heard things about active appearance models but I'm having trouble understanding how to use this with python and I don't have enough context to figure out if this is even the conventional method for detecting different parts of the lips. Is that my best choice or do I have other options? If I should use it, how would I get started with it using python and simplecv?
回答1:
Here are two papers for lip segmentation using edge detection and color segmentation. Try them out.
paper1, paper2
回答2:
The color segementation involves "gradient of the difference between the pseudo-hue and luminance (obtaining hybrid contours)". Try googling for qouted string and you will find multiple research papers on this topic.
来源：https://stackoverflow.com/questions/8840127/opencv-lip-segmentation