Video decoding method
AssigneeHitachi Consumer Electronics Co., Ltd.
InventorsShohei Saito, Masashi Takahashi, Muneaki Yamaguchi, Hiroaki Ito, Koichi Hamada
Life of this patent
Estimating the term…
Abstract
A method of decoding videos comprises the steps of performing, in respect of a decoding objective area of decoding objective frame, a motion search by using images of plural frames finished with decoding and deciding, on the basis of the result of the motion search, whether an image of the decoding objective area is to be generated through an interpolation process or a decoded image is to be generated through motion compensation using data included in an encoding stream.
Claims20
A video decoding method of decoding an encoded video stream comprising the steps of:
- receiving a motion search direction decision flag in the encoded video stream;
- searching, with respect to a decoding objective area of a decoding objective frame, for a motion vector according to a motion search direction specified by the motion search direction decision flag using a plurality of decoded images of a plurality of decoded frames where the decoding objective frame is an interpolation frame positioned where two of the plurality of decoded frames are linearly connected with the decoding objective area, and an objective pixel included in said decoding objective area is positioned between and on a same straight line in a frame of space and time as each of plural pixels positioned on said decoded images of plural decoded frames;
- and
- deciding, on the basis of a result of said motion vector search, whether an image of said decoding objective area is to be generated through an interpolation process in accordance with the motion search direction or a decoded image is to be generated through motion compensation using data included in the encoded video stream.
A video decoding method of decoding an encoded video stream comprising the steps of:
- deciding, on the basis of a mode decision flag included in the encoded video stream, whether an image of a decoding objective area is to be generated through an interpolation process using a plurality of decoded images of a plurality of decoded frames of the encoded video stream where the decoding objective area belongs to a decoding objective frame which is an interpolation frame positioned where two of the plurality of decoded frames are linearly connected with the decoding objective area or through a motion compensation process using data included in the encoded video stream, wherein the interpolation process is performed under a condition that an objective pixel included in said decoding objective area is positioned between and on a same straight line in a frame of space and time as each of plural pixels positioned on said decoded images of plural decoded frames;
- and
- generating a decoded image by switching over, in accordance with the result of decision in said decision step, from an interpolation process using decoded image to a motion compensation process using data included in the encoded video stream and vice versa,
- wherein the image of the decoding objective area is generated through an interpolation process according to a motion search direction decision flag included in the encoded video stream which specifies an interpolation direction in which the image is generated.
A video decoding method of decoding an encoded video stream including a first flag indicative of the number of consecutive image blocks to be generated through an interpolation process using decoded images in a decoding objective area and a second flag indicating that a decoded image is to be generated with respect to an image block of the decoding objective area through a motion compensation process using data included in the encoded video stream, comprising the steps of:
- detecting a flag from said encoded video stream to decide whether the detected flag is the first flag or the second flag;
- and
- generating a decoded image by switching over, in accordance with the decision result in said detecting step, an interpolation process using a plurality of decoded images of a plurality of decoded frames where the consecutive image blocks are interpolated at a position where two of the plurality of decoded frames are linearly connected with the consecutive image blocks and applied to plural image blocks corresponding to the first flag to a motion compensation process using data included in the encoded video stream and applied to an image block corresponding to the second flag, wherein the interpolation process is performed under a condition that an objective pixel included in said decoding objective area is positioned between and on a same straight line in a frame of space and time as each of plural pixels positioned on said decoded images of plural decoded frames,
- wherein in the interpolation process, the consecutive image blocks are interpolated according to a motion search direction decision flag included in the encoded video stream which specifies an interpolation direction.
Dependencies are inferred from the claim text.