University of California

Mainak Biswas

Truong Q Nguyen

Fabrizio Vacca

Andrea Molino

Sanjeev Kumar

A video codec of a preferred embodiment is suitable for implementation on a variety of platforms, including platforms where devices require both encoding and decoding engines. Various features of preferred embodiment codecs reduce complexity for operations such as motion determination. Channel resources are allocated in preferred embodiments unequally, with an allocation being determined based upon channel conditions and the error level in a frame being encoded. A decoder of a preferred embodiment, provided with an error frame and motion vector information may conduct decoding with motion compensation. In an example encoder of the invention, a motion estimator identifies motion vectors using a phase plane correlation technique based on phase quantization, where the phase difference between the Fourier transformed version of blocks in a reference frame and a current frame is estimated by dividing the complex plane into sectors and assigning the Fourier domain difference to the sectors.

Video encoding methods and devices

The invention provides methods for global motion estimation, determining a coarse estimation, and refining a coarse estimation. Embodiments of the invention provide a fast and robust global motion estimation algorithm based on two-stage coarse-to-fine refinement strategy, which is capable of measuring large motions. An embodiment of the invention may be applied as a modification of any standard, e.g. MPEG-4 that uses the affine model of motion estimation. Embodiments of the invention may be used in the six parameter affine motion model, and other embodiments of the invention are applicable to the two parameter translation model, the four parameter RST model, and the eight parameter projective model. In a preferred embodiment, a coarse estimation is developed in a translation invariant domain, and then is refined in the spatial domain.

Global motion estimation image coding and processing

A phase plane correlation motion vector determination method identifies images areas in which motion took place. A reference image is shifted by the candidate motion vectors obtained from the phase plane correlation. The shifted reference image is correlated to determine similarity with the current image and to eliminate spurious motion vectors. Candidate motion vectors are validated to determine valid motion vectors and identify problem areas. In a preferred embodiment, a correlation resolution is set to a sub-block size. Sub-blocks of the reference image are shifted by candidate motion vectors for corresponding blocks. Shifted sub-blocks are correlated with sub-blocks of a current image using a sub-block neighborhood to determine valid vector assignments. Where an assignment proves invalid, global motion parameters are determined to assign motion vectors to sub-blocks where correlation fails.

Growing community of inventors

La Jolla, CA, United States of America

Mainak Biswas