Bidirectional predicted pictures or video object planes for efficient and flexible coding

A decoder that decodes a block in a group of blocks, the group of blocks associated with a partition of a picture, at least some blocks of the group of blocks being B type blocks, the decoder comprising:

• a module configured to decode information related to a shape associated with the group of blocks;
• a module configured to receive an indication that a prediction mode associated with a block in the group of blocks is a direct prediction mode, the prediction mode being one of four possible prediction modes including a direct prediction mode, a forward prediction mode, a backward prediction mode and an interpolation prediction mode;
• a module configured, based at least on the received indication of prediction mode for the block, to derive a motion vector from a temporally future picture;
• and
• a module configured to predict data for the block using the derived motion vector.

A tangible computer-readable medium storing instructions for controlling a device to decode a block in a group of blocks, the group of blocks associated with a partition of a picture, at least some blocks of the group of blocks being B type blocks, the instructions comprising:

• decoding information related to a shape associated with the group of blocks;
• receiving an indication that a prediction mode associated with a block in the group of blocks is a direct prediction mode, the prediction mode being one of four possible prediction modes including a direct prediction mode, a forward prediction mode, a backward prediction mode and an interpolation prediction mode;
• based at least on the received indication of prediction mode for the block, deriving a motion vector from a temporally future picture;
• and
• predicting data for the block using the derived motion vector.