Method and apparatus of motion data buffer reduction for three-dimensional video coding

A method for three-dimensional video coding, multi-view video coding and scalable video coding, the method comprising:

• receiving input data associated with a picture unit in an access unit, wherein the picture unit corresponds to a texture picture or depth map, or a slice of the texture picture or depth map, and the picture unit is divided into 4×4 blocks;
• applying first 4 to 1 compression to first motion data associated with the picture unit in the access unit by using a top-left 4×4 block of an 8×8 block as a first representative block, wherein said applying first 4 to 1 compression is performed within a substantially short period after the picture unit in the access unit is coded or decoded;
• storing motion data of the first representative blocks associated with the picture unit in a motion data buffer, wherein all 4×4 blocks within the 8×8 block share same first motion data stored for the first representative block during first encoding or decoding process of another picture unit in the access unit;
• and
• applying second 4 to 1 compression to the motion data associated with the picture units in the access unit by using the top-left 4×4 block of a 16×16 block as a second representative block, wherein said applying second 4 to 1 compression is performed after all picture units in the access unit are coded or decoded, wherein all 4×4 blocks within the 16×16 block share same second motion data stored for the second representative block during second encoding or decoding process of the picture units in another access unit.

An apparatus for three-dimensional video coding, multi-view video coding and scalable video coding, the apparatus comprising:

• one or more electronic circuits, wherein said one or more electronic circuits are configured to:
• receive input data associated with a picture unit in an access unit, wherein the picture unit corresponds to a texture picture or depth map, or a slice of the texture picture or depth map, and the picture unit is divided into smallest motion data blocks, and a motion sharing area consisting of a first plurality of smallest motion data blocks is selected for motion data sharing;
• divide the motion sharing area into multiple to-be-compressed units, wherein each to-be-compressed unit consists of a second plurality of smallest motion data blocks;
• select second representative motion data for the second plurality of smallest motion data blocks to share;
• store first compressed motion data corresponding to the second representative motion data associated with the picture unit in a buffer after the picture unit is encoded or decoded, wherein the second representative motion data associated with the picture unit is used for first encoding or decoding process of another picture unit in the access unit;
• select first representative motion data for the first plurality of smallest motion data blocks to share;
• and
• generate second compressed motion data based on the first representative motion data associated with the picture units of the access unit, wherein the second compressed motion data is used for second encoding or decoding process of the picture units in another access unit.