Action Recognition by Dense Trajectories

INTRODUCTION

in general，trajectory features is extracted using the KLT tracker or matching SIFT descriptors between frames.
however，extracted trajectores is various depend on the methods， We sample dense points from each frame and track them based on displacement information from a dense optical flow field.

The trajectories are obtained by tracking ensely sampled points using optical flow fields. The number
of tracked points can be scaled up easily, as dense flow
fields are already computed. Furthermore, global smoothness
constraints are imposed among the points in dense optical
flow fields, which results in more robust trajectories than
tracking or matching points separately