Image encoding and decoding method and apparatus using motion compensation filtering
AssigneeSamsung Electronics Co., Ltd.
InventorsKyo-hyuk Lee, Mathew Manu
Life of this patent
Estimating the term…
Abstract
Provided are an image encoding method and apparatus for encoding image data, in which compression efficiency is further improved by filtering motion-compensated image data, thereby reducing errors with respect to original image data, and an image decoding method and apparatus for decoding encoded image data. A filter is generated using previously decoded surrounding pixels of a current frame and surrounding pixels of a reference frame, and prediction data is then filtered using the generated filter. Therefore, it is possible to decrease a size of residual data to be encoded.
Claims19
An image encoding method comprising:
- using a processor to perform a method including:
- generating a motion vector of a current block to be encoded using motion prediction of the current block;
- generating a prediction block for the current block using motion compensation for obtaining a corresponding block of a reference frame indicated by the motion vector;
- filtering the prediction block generated using the motion compensation;
- and
- encoding a difference between the filtered prediction block and the current block,
- wherein the filtering the prediction block comprises:
- generating a filter used for the prediction block by using a correlation between pixels surrounding the current block that have been previously encoded and restored, and pixels surrounding the corresponding block of the reference frame;
- and
- filtering respective pixel values of the prediction block using the generated filter.
An image encoding apparatus comprising:
- a motion predictor that generates a motion vector of a current block to be encoded through motion prediction of the current block;
- a motion compensator that generates a prediction block for the current block using motion compensation, by obtaining a corresponding block of a reference frame indicated by the motion vector;
- a filtering unit that filters the prediction block generated by the motion compensator;
- and
- an encoder that encodes a difference between the filtered prediction block and the current block,
- wherein one or more of the motion predictor, motion compensator, filtering unit, and encoder are implemented on a processor, and
- the filtering unit generates a filter for filtering the generated prediction block by using a correlation between pixels surrounding the current block that have been previously encoded and restored, and pixels surrounding the corresponding block of the reference frame;
- and filters respective pixel values of the generated prediction block using the generated filter.
An image decoding method comprising:
- using a processor to perform a method including:
- generating a prediction block for a current block to be decoded through motion compensation using a motion vector extracted from a received bit-stream;
- generating a filter to be used for the prediction block using a correlation between pixels surrounding the current block and pixels surrounding a corresponding block of a reference frame indicated by the motion vector;
- generating a filtered prediction block by filtering the prediction block using the filter;
- and
- restoring the current block by adding the filtered prediction block and a restored residual,
- wherein, the generating of a filter comprises determining weight values of the filter such that an error is minimized between pixels surrounding the corresponding block of the reference frame and pixels surrounding the current block, after the pixels surrounding the corresponding block of the reference frame are filtered using the filter.
An image decoding apparatus comprising:
- a motion compensator that generates a prediction block for a current block to be decoded through motion compensation using a motion vector extracted from a received bit-stream;
- a filtering unit that generates a filter to be used for the prediction block using a correlation between pixels surrounding the current block and pixels surrounding a corresponding block of a reference frame indicated by the motion vector, the filtering unit generating a filtered prediction block by filtering the generated prediction block using the filter;
- and
- an adder that restores the current block by adding the filtered prediction block and a restored residual,
- wherein one or more of the motion compensator, filtering unit, and adder are implemented on a processor, and
- the filtering unit determines weight values of the filter such that an error is minimized between pixels surrounding the corresponding block of the reference frame and pixels surrounding the current block, after the surrounding pixels of the corresponding block of the reference frame are filtered using the filter.
Dependencies are inferred from the claim text.