Method and apparatus for encoding image by using atypical split, and method and apparatus for decoding image by using atypical split
AssigneesSamsung Electronics Co., Ltd., Industry-Academic Cooperation Foundation, Yonsei University
InventorsSun-il Lee, Won Woo Ro, Kyungah Kim, Minsik Kim, Chan-yul Kim
Life of this patent
Estimating the term…
Abstract
Provided is an image encoding method including extracting feature points from a picture; generating at least two clusters by performing feature point clustering on the extracted feature points; determining at least two split sections in the picture, the at least two split sections respectively corresponding to the at least two clusters; parallel-encoding the at least two split sections; and generating a bitstream including information about the at least two split sections. A size and a shape of each of the at least two split sections may be individually determined.
Claims17
An image encoding method comprising:
- extracting feature points from a picture;
- generating at least two clusters by performing feature point clustering on the extracted feature points;
- determining at least two split sections in the picture, the at least two split sections respectively corresponding to the at least two clusters, wherein each split section is configured of at least one block;
- parallel-encoding the at least two split sections;
- and
- generating a bitstream comprising information about the at least two split sections,
- wherein the determining the at least two split sections comprises, when a largest coding unit (LCU) includes the feature points of the at least two clusters, determining the LCU to be included in a specific split section from among the at least two split sections, the specific split section corresponding to a cluster having a highest number of feature points from among the at least two clusters, and
- wherein a size and a shape of each of the at least two split sections are individually determined.
An image encoding apparatus comprising:
- a cluster generator configured to extract feature points from a picture, and generate at least two clusters by performing feature point clustering on the extracted feature points;
- a split section determiner configured to determine at least two split sections in the picture, the at least two split sections respectively corresponding to the at least two clusters, wherein each split section is configured of at least one block;
- and
- a bitstream generator configured to parallel-encode the at least two split sections, and generate a bitstream comprising information about the at least two split sections,
- wherein the split section determiner is further configured to, when a largest coding unit (LCU) includes the feature points of the at least two clusters, determine the LCU to be included in a specific split section from among the at least two split sections, the specific split section corresponding to a cluster having a highest number of feature points from among the at least two clusters, and
- wherein a size and a shape of each of the at least two split sections are individually determined.
An image decoding method comprising:
- receiving a bitstream comprising split section information;
- determining at least two split sections based on the split section information, wherein each split section is configured of at least one block;
- and
- reconstructing an image by performing parallel-decoding on the at least two split sections.
An image decoding apparatus comprising:
- a receiver configured to receive a bitstream comprising split section information;
- a split section determiner configured to determine at least two split sections based on the split section information, wherein each split section is configured of at least one block;
- and
- a decoder configured to reconstruct an image by performing parallel-decoding on the at least two split sections.
Dependencies are inferred from the claim text.