Facial landmark detection and tracking with dynamically adaptive matched filters
What's it about?
Landmarks are salient features that describe shape, color, or texture throughout simple data. Facial landmark detection is essential in face recognition, pose estimation, and facial expression recognition. The existing approaches strongly depends on performing an intensive training stage using an appropriate set of training images.
Why is it important?
A reliable method for detection and tracking of facial landmarks is presented. A bank of composite matched filters is constructed given a set of prespecified facial landmarks in a reference face image. The filter bank is dynamically adapted to each captured frame by learning from current and past detections considering geometrical modifications.
The proposed method can be helpful in several applications such as face recognition, pose estimation, and facial expression recognition.