Multi-Layer Video DPB Handling for IDR Reference Preservation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Managing a decoded picture buffer (DPB) for multi-layer video data encoding and decoding is challenging due to the need to determine which pictures to keep or remove, especially when an instantaneous decoding refresh (IDR) picture is encountered, as it can lead to non-decodable bitstreams due to unavailable reference pictures.

Innovation Solution

A video decoder removes all decoded pictures of the layer containing an IDR picture from the DPB while retaining those with lower layer IDs, ensuring future pictures can be decoded without the IDR picture, thus maintaining a decodable bitstream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If all decoded pictures are removed from the DPB when an IDR picture is encountered, then the DPB is cleared for new reference pictures, but the bitstream becomes non-decodable due to unavailable reference pictures

Engineering Contradiction:
Improvedecodability of bitstreamVSAvoidDPB management complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the DPB management by layer ID, treating different layers independently. When an IDR picture is encountered in a particular layer, only pictures in that specific layer are removed from the DPB, while pictures in other layers are preserved. This segmentation allows the system to maintain reference pictures across layers, ensuring the bitstream remains decodable while still clearing the DPB appropriately for the current layer.

Inventive Principle:
Principle #1Segmentation

2Reliability

If pictures from higher layers are removed when an IDR picture is encountered, then the DPB is prepared for new references, but necessary reference pictures become unavailable

Engineering Contradiction:
Improveavailability of reference picturesVSAvoiddecoding efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies segmentation by layer ID to preserve reference pictures across layers. When an IDR picture is encountered in layer N, only pictures in layer N are removed from the DPB, while pictures in layers other than N are preserved. This ensures that reference pictures remain available for decoding operations, maintaining both reliability and productivity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements preliminary action by checking the layer ID of the current picture before removing pictures from the DPB. This preliminary check ensures that only pictures from the correct layer are removed, preventing the loss of necessary reference pictures from other layers that are still needed for decoding.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If the DPB is not cleared when an IDR picture is encountered, then reference pictures remain available, but the DPB contains outdated pictures that may cause decoding errors

Engineering Contradiction:
Improvedecoding accuracyVSAvoidnumber of pictures in DPB
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the DPB by layer ID, removing only pictures from the specific layer containing the IDR picture while preserving pictures from other layers. This segmented approach ensures that outdated pictures are removed from the appropriate layer without inadvertently removing necessary reference pictures from other layers, maintaining decoding accuracy while managing the quantity of pictures in the DPB.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3981154B1Spatial scalability support in video encoding and decoding
Publication Date: 2025.11.12 QUALCOMM INC
  • EP3981154B1 patent drawingFigure 1
  • EP3981154B1 patent drawingFigure 2A~2B
  • EP3981154B1 patent drawingFigure 3

AI summary

A video encoder or video decoder can be configured to obtain multi-layer video data that includes at least a first layer and a second layer; determine a layer identification value for the first layer and a layer identification value for the second layer, wherein the layer identification value for the second layer is higher than the layer identification value for the first layer; store decoded pictures of the first layer and decoded pictures of the second layer in a decoded picture buffer; and in response to determining that a picture of a layer is an instantaneous decoding refresh (IDR) picture, remove from the decoded picture buffer, all decoded pictures of a layer the IDR picture belongs to, while leaving in the decoded picture buffer all decoded pictures with layer identifications lower than a layer identification of the layer the IDR picture belongs to.