False matching due to errors in feature extraction and changes in illumination between frames may occur in feature tracking in image sequences. False matching leads to outliers in...
Varying illumination and partial occlusion are two main difficulties in visual tracking. Existing methods based on appearance information cannot solve these problems effectively s...
We present an algorithm that learns invariant features from real data in an entirely unsupervised fashion. The principal benefit of our method is that it can be applied without hu...
3D tracking of faces in video streams is a difficult problem that can be assisted with the use of a priori knowledge of the structure and appearance of the subject’s face at pred...