Video and Image NAL Header Extensions for Layer IDs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as HEVC and VVC, are limited by the number of layer ID values in NAL unit headers, which restricts flexibility and efficiency in handling multiple layers, and do not support dependent coding between layers.
Innovation Solution
An extension mechanism is introduced for the NAL unit headers to increase the number of layer ID values by using an optional extension value in the header, allowing the layer ID to be generated from a combination of the initial field value and an additional extension field, using formulas like Layer ID = V0 + (F - V0) * 2^n + S or Layer ID = V0 + ((F - V0) << n) + S, where F is the initial field value, V0 is the start value, n is the number of bits for the extension, and S is the extension value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the NAL unit header uses a fixed number of bits for layer ID, then the header structure is simple and decoding is fast, but the number of supported layer ID values is limited
Solution Approach 1:
The patent applies dynamics by making the NAL unit header structure adaptive rather than fixed. The header can dynamically switch between a standard format (for regular cases) and an extended format (when layer ID exceeds the standard range). This is achieved through a syntax element that indicates whether the extended format is used, allowing the header to adjust its structure based on the actual data requirements, thereby supporting more layer ID values without permanently increasing complexity for all cases.
Solution Approach 2:
The patent changes the parameter representation by introducing an extended syntax element that modifies how layer ID is encoded. Instead of using a fixed number of bits, the system transitions to a different encoding scheme when the layer ID value exceeds the standard range. This parameter change allows the header to represent larger layer ID values while maintaining backward compatibility with existing decoders that can recognize and handle both the standard and extended formats.
2Adaptability or versatility
If the NAL unit header is extended to support more layer ID values, then flexibility in handling multiple layers is improved, but the header size and processing complexity increase
Solution Approach 1:
The patent applies partial action by extending the header functionality only when necessary. The extended syntax element and additional fields are introduced conditionally - only when the layer ID exceeds the standard range. For the majority of cases where the standard layer ID range is sufficient, the header remains unchanged and decoding proceeds at normal speed. This selective extension minimizes the impact on overall decoding efficiency while providing the needed flexibility for complex multi-layer scenarios.
3Productivity
If dependent coding between layers is supported, then video compression efficiency is improved, but the complexity of encoding and decoding processes increases
Solution Approach 1:
The patent introduces an intermediary mechanism through the extended NAL unit header that facilitates dependent coding between layers. The header contains syntax elements that explicitly indicate layer dependencies and relationships, acting as a mediator that allows the decoder to efficiently manage inter-layer predictions and references. This structured approach to representing layer relationships enables compression efficiency gains from dependent coding while keeping the complexity manageable through standardized syntax elements that can be systematically processed.
Data Source
AI summary
A method of decoding an encoded image includes obtaining a layer identification, Layer ID, indicator from a first field of a header for a network abstraction layer, NAL, unit of the encoded image, determining whether the Layer ID indicator is less than a start value, in response to determining that the Layer ID indicator is not less than the start value, determining the Layer ID value based on the Layer ID indicator and based on an extension value provided in a second field of the header, and decoding the NAL unit based on the Layer ID value. Related decoders, encoding methods, encoders and computer program products are disclosed.


