# US 9,438,907 B2 · Motion picture encoding apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 9,438,907 B2
- Document ID: US9438907B2
- Status: Granted
- Kind: B2
- Application number: 13/983,021
- Filed: 2011-11-22
- Published: 2013-11-21
- Granted: 2016-09-06
- Earliest claimed: 2011-02-17
- Assignees: Hitachi Kokusai Electric Inc.
- Applicants: Tadakazu Kadoto; Masashi Takahashi; Nobuhiro Chihara
- Inventors: Tadakazu Kadoto; Masashi Takahashi; Nobuhiro Chihara
- Web page: https://patentgrep.com/patent/US9438907B2

## CPC

- H04N 19/50: using predictive coding (H04N19/61 takes precedence)
- H04N 19/11: among a plurality of spatial predictive coding modes
- H04N 19/136: Incoming video signal characteristics or properties
- H04N 19/176: the region being a block, e.g. a macroblock
- H04N 19/61: in combination with predictive coding
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

Suppression of prediction error accumulating in an intra prediction is described. A motion picture encoding apparatus that performs encoding processing for each pixel block using a motion picture encoding method includes intra prediction processing for generating a prediction image using correlation of pixels therearound, and in the intra prediction processing, the motion picture encoding apparatus has processing for generating a quasi-prediction image from an input image, wherein when a prediction mode is determined, a weight is given to a pixel serving as a reference pixel in the intra prediction processing, and thus, the prediction mode is determined. When the weight is set at one or more, an error of a pixel serving as a reference pixel in a subsequent block is emphasized, and therefore, a prediction mode that is likely to propagate the error to the subsequent block is less likely to be selected.

## Claims

### Claim 1

A motion picture encoding method that encodes a pixel block upon selecting one of a plurality of intra prediction modes, the motion picture encoding method comprising:
a step of generating a plurality of quasi-prediction images according to the plurality of intra prediction modes from an input image;
a step of calculating a plurality of prediction errors of the plurality of prediction images generated by a quasi-prediction image generating unit by giving a weight emphasizing a prediction error in a pixel that could be a reference pixel in an intra prediction of another pixel block and determining a prediction mode on a basis of the plurality of calculated prediction errors;
a step of generating a prediction image from a reconfigured image in accordance with the determined prediction mode;
a step of obtaining a difference image by subtracting the prediction image generated from the reconfigured image;
a step of obtaining a coefficient by Performing orthogonal transformation of the difference image and quantizing the coefficient; and
a step of performing dequantization and inverse orthogonal transformation of the quantized coefficient and updating the reconfigured image on a basis of an image obtained by adding the coefficient to the prediction image generated from the reconfigured image;
wherein in the step of calculating, the plurality of calculated prediction errors for the pixel block are calculated by a formula:
SAD
′
=
∑
i
⁢

Input
⁡
(
i
)
-
Pred
⁡
(
i
)

+
δ
⁢
∑
j
⁢

Input
⁡
(
j
)
-
Pred
⁡
(
j
)

,
wherein SAD′ is prediction errors, Input(i) and Input(j) are pixels of the input image, Pred(i) and Pred(j) are pixels of a quasi-prediction image, i is a suffix for pixels other than a weighting target pixel, j is a suffix for the weighting target pixel, and δ denotes the weight.