Method and apparatus for encoding and decoding a video signal using feature point based motion estimation

An apparatus, for use in an image signal encoding system, for encoding a bidirectionally predictive frame (B frame) of a digital video signal based on two reconstructed neighboring frames of the B frame by employing a feature point based motion estimation approach to generate an encoded B frame signal, which comprises the steps of:

• means for providing a first reconstructed neighboring frame as a first reference frame and providing a second reconstructed neighboring frame as a second reference frame;
• a detection means for detecting a processing region representing a motion of a moving object of the digital video signal based on a difference between the first and second reconstructed neighboring frames to generate region information denoting the detected processing region;
• a first selection means for selecting two groups of pixels from pixels contained in the detected processing regions of the first and the second reference frames as a first and a second sets of feature points based on the region information, respectively;
• a motion estimation means for determining a first set of motion vectors between the current and the first reference frames and determining a second set of motion vectors between the current and the second reference frames, each of the of motion vectors representing a motion for each of the feature points;
• a second selection means for selecting the set of motion vectors for the set of feature points from the first and the second sets of motion vectors to thereby generate a selected set of motion vectors and selection information representing the selected set as the encoded B frame.

A method, for use in an image signal decoding system, for decoding an encoded bidirectionally predictive frame (encoded B frame) of an encoded digital video signal based on two reconstructed neighboring frames of the encoded B frame by employing a feature point based motion estimation approach to generate a reconstructed B frame signal, wherein the B frame includes a set of motion vectors and selection information, which comprises the steps of:

• (a) selecting one of reconstructed neighboring frame as a reference frame based on the selection information;
• (b) detecting a processing region representing a motion of a moving object of the digital video signal based on a difference between the reconstructed neighboring frames to generate region information denoting the detected processing region;
• (c) selecting a number of pixels from pixels contained in the detected processing region of the reference frame as a set of feature points based on the region information;
• and
• (d) providing the reconstructed B frame by using the set of motion vectors, the selected set of feature points and the reference frame.

An apparatus, for use in an image signal decoding system, for decoding an encoded bidirectionally predictive frame (encoded B frame) of an encoded digital video signal based on two reconstructed neighboring frames of the encoded B frame by employing a feature point based motion estimation approach to generate a reconstructed B frame signal, wherein the B frame includes a set of motion vectors and selection information, which comprises:

• a first selection means for selecting one of reconstructed neighboring frames as a reference frame based on the selection information;
• a detection means for detecting a processing region representing a motion of a moving object of the digital video signal based on a difference between the reconstructed neighboring frames to generate region information denoting the detected processing region;
• a second selection means for selecting a number of pixels from pixels contained in the detected processing region of the reference frame as a set of feature points based on the region information;
• and
• a generation means for providing the reconstructed B frame by using the set of motion vectors, the selected set of feature points and the reference frame.