Patent
Moving image coding/decoding device using moving compensation inter-frame prediction system employing affine transformation
Ema Nobuyuki,Hibi Keiichi +1 more
Reads0
Chats0
TLDR
In this paper, the authors proposed a method to conduct moving compensation inter-frame prediction by obtaining a motion vector between image frames with high accuracy, conducting optimum area division to an object shape or the like and applying efficient coding/decoding to the motion vector.Abstract:
PROBLEM TO BE SOLVED: To conduct moving compensation inter-frame prediction by obtaining a motion vector between image frames with high accuracy, conducting optimum area division to an object shape or the like and applying efficient coding/decoding to the motion vector. SOLUTION: A moving compensation inter-frame prediction section 17 changes a coding processing area based on an input image and a reference image in a frame memory section 16 to generate a predicted image and to obtain coding side information. The prediction section 17 selects a valid/invalid mask for processing from a position of a processing object at the retrieval by a vector retrieval section 32 by using a valid area selection section 31, divides the processing area suitable for the image from a motion vector and the reference image and uses affine transformation and parallel displacement to generate a predicted image by using a variable area size prediction frame generating section 33, controls the area division and allows an area division pattern decision section 34 to provide an output of side information such as the predicted image, the motion vector and area division information. A side information coding section 35 codes an additional motion vector as a difference from a mean of basic motion vectors or applies prediction coding to the motion vector from a median of three adjacent motion vectors.read more
Citations
More filters
Patent
Inter-prediction mode based image processing method, and apparatus therefor
Jaehyun Lim,Naeri Park +1 more
TL;DR: In this paper, an inter-prediction mode based image processing method and an apparatus therefor are disclosed, which may comprise the steps of: deriving movement information of a control point for specifying a reference block of a current block, dividing the current block into sub-blocks, deriving the movement information for each sub-block using the control point, and generating a prediction block for each of the subblocks.
Patent
Inter-prediction method and apparatus in video coding system
TL;DR: In this article, an inter-prediction method performed by a decoding apparatus according to the present invention comprises the steps of: deriving a motion information candidate list of a current block on the basis of neighboring blocks of the current block; selecting a specific candidate on basis of the comparison of the candidates in the motion information candidates list; and generating a prediction sample of the predicted motion vector of the target block based on the current vector of a target block.
Patent
Intra prediction method and device in video coding system
TL;DR: In this article, a prediction sample can be adaptively filtered according to the intra prediction mode, and intra prediction performance can be improved by applying filtering to the prediction sample on the basis of the filtering reference samples.
Patent
Image encoding device, image encoding method, image encoding program, image decoding device, image decoding method, and image decoding program
TL;DR: In this paper, a residual approximation signal creation unit (RAS) is used to create a residual signal by linearly approximating the first residual signal on the basis of the direct-current component of the residual signal and a subtraction unit (307) calculates a second residual signal from a difference between the original signal of the image and the in-plane prediction signal.
Patent
Encoding device, decoding device, encoding method, and decoding method
TL;DR: In this paper, an encoding device (100) consisting of a circuit (160) and memory (162) is used to perform a step in which an image block is partitioned into a plurality of partitions, and then the prediction images for the partitions are used to encode the image block.