Hello everybody,

I am processing images of people and I am trying to find out if the person is smiling or not. My idea is to find the contour of the lips and implement some kind of machine learning to determine if is is a smile or not.

The setup:

• I am using Emgu CV library
• The images are grayscaled and static

What do I do now:

• find person's face using Cascase classifier
• find the person's mouth within the face using Cascase classifier
• find the ALL contours in this mouth rectangle(one of the all should be lips contour)

The result:

• it returns only one contour that is the boundary of the mouth rectangle which is NOT the desired result.

I am new to Emgu CV library and still reading through its documentation and examples, but if you have some experience with it or with computer vision in general, please share your thoughts.

Have a nice day!