Picture prediction method and apparatus

A method comprising:

• encoding an identifier into a bitstream, wherein the identifier indicates whether a combined intra and inter prediction mode is applied for a current block;
• determining that a geometric partitioning based motion compensation is used to obtain prediction samples of the current block when the identifier indicates that the combined intra and inter prediction mode is not applied for the current block;
• obtaining a splitting direction based on a geometric partitioning index when the geometric partitioning based motion compensation is used;
• and
• encoding the geometric partitioning index into the bitstream when the geometric partitioning based motion compensation is used.

A device comprising:

• a receiver configured to receive one or more bitstreams, wherein each of the one or more bitstreams comprises encoded picture data and an identifier indicating whether a combined intra and inter prediction mode is applied for a current block, wherein a geometric partitioning based motion compensation is used to obtain prediction samples of the current block when the identifier indicates that the combined intra and inter prediction mode is not applied for the current block, and wherein when the geometric partitioning based motion compensation is used, the one or more bitstreams further comprise a geometric partitioning index and a splitting direction is based on the geometric partitioning index;
• a memory configured to store the one or more bitstreams;
• and
• one or more processors coupled to the memory and configured to reconstruct a picture from the one or more bitstreams stored in the memory using the identifier and the geometric partitioning index.