Reduced Memory Video Decode via Decimation and Interpolation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video decoding methods are memory-intensive, requiring significant storage resources and bandwidth due to full-resolution decoding, which is inefficient in terms of storage and memory usage.

Innovation Solution

The implementation of a modified MPEG-2 decoder with a filtering and decimation unit, pixel interpolation unit, and conditional pre-processing unit that reduces memory requirements by filtering and decimating macroblocks, interpolating pixels, and conditionally pre-processing reference blocks to generate a decimated video stream, which reduces storage needs and enhances decoding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If full-resolution decoding is performed, then picture quality is maintained, but memory requirements and storage resources increase significantly

Engineering Contradiction:
Improvepicture qualityVSAvoidmemory requirements
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The video stream is divided into macroblocks that are processed independently at reduced resolution. The filtering and decimation unit segments the video data into manageable units, allowing memory-efficient processing while maintaining quality through selective reconstruction of high-resolution regions only when necessary.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of decoding the entire video stream at full resolution, the system applies partial action by decoding at reduced resolution and selectively enhancing only specific regions or frames where high quality is critical. This reduces overall memory usage while maintaining perceptual picture quality.

Inventive Principle:
Principle #16Partial or excessive action

2Measurement precision

If full-resolution decoding is performed, then picture quality is maintained, but data transfer bandwidth increases

Engineering Contradiction:
Improvepicture qualityVSAvoiddata transfer bandwidth
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The invention extracts only the essential visual information at reduced resolution for initial decoding and playback. High-resolution data is extracted and transmitted only when and where needed, separating the baseline quality stream from enhancement data to reduce overall bandwidth requirements.

Inventive Principle:
Principle #2Taking out (Extraction)

3Quantity of substance

If reduced resolution decoding is used, then memory requirements decrease, but picture quality deteriorates

Engineering Contradiction:
Improvememory requirementsVSAvoidpicture quality
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The system dynamically changes resolution parameters based on content importance, motion activity, and scene complexity. Critical regions maintain higher resolution while less important areas use lower resolution, optimizing the balance between memory usage and perceived picture quality through adaptive parameter adjustment.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS9326004B2Reduced memory mode video decode
Publication Date: 2016.04.26 AVAGO TECHNOLOGIES INTERNATIONAL SALES PTE LTD
  • US9326004B2 patent drawing
  • US9326004B2 patent drawing
  • US9326004B2 patent drawing

AI summary

A method and system to decode a video stream are provided. The method comprises receiving macroblocks, filtering and decimating the macroblocks to create decimated macroblocks and storing the decimated macroblocks. The method further comprises creating a decimated reference block from one or more decimated macroblocks of a decimated reference picture and interpolating selected pixels of the decimated reference block to create an interpolated reference block. The method further comprises pre-processing selected columns of the interpolated reference block to create a processed reference block for motion compensation.