Image Bitstream Decoding With NNPF SEI Picture-Type Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images leads to a significant increase in transmission and storage costs due to the rise in the amount of transmitted information, necessitating high-efficient image compression technology, and the need to process neural-network post-filter (NNPF)-related supplemental enhancement information (SEI) messages without errors, while restricting certain unit information to avoid mismatches between encoders and decoders.
Innovation Solution
An image encoding/decoding method that restricts the type of current pictures based on NNPF-related SEI messages, ensuring that discardable and non-output pictures are not included in the unit information, thereby preventing mismatches between the encoder and decoder, and includes a method for transmitting and storing bitstreams generated by this process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images are transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent changes the parameter of image resolution and quality to provide high-definition and ultra-high-definition images, which directly leads to increased amount of transmitted information and higher transmission costs. This is the core technical contradiction that the patent aims to resolve through neural-network post-filtering techniques.
2Adaptability or versatility
If discardable pictures are included in NNPF input picture list, then encoder flexibility is improved, but decoder reliability deteriorates due to mismatch errors
Solution Approach 1:
The patent applies preliminary action by having the encoder explicitly signal the input picture list for the neural-network post-filter to the decoder before decoding occurs. This allows the decoder to prepare and avoid mismatches, ensuring that both encoder and decoder use the same picture list without including discardable pictures, thereby maintaining reliability while preserving flexibility.
Solution Approach 2:
The patent implements feedback by transmitting information about the NNPF input picture list from the encoder to the decoder through the bitstream. This feedback mechanism ensures that the decoder has accurate knowledge of which pictures should be used as inputs, preventing reliability issues caused by mismatches while allowing the encoder to maintain flexibility in picture selection.
3Adaptability or versatility
If NNPFC SEI messages are associated with all picture types, then message coverage is improved, but processing complexity increases due to error-prone operations
Solution Approach 1:
The patent applies local quality by restricting NNPFC SEI messages to be associated only with non-discardable pictures (such as output pictures and reference pictures) rather than all picture types. This localized approach reduces processing complexity and error-prone operations while maintaining adequate message coverage for pictures that actually need neural-network filtering.
Solution Approach 2:
The patent uses partial action by selectively applying NNPFC SEI messages only to the necessary subset of pictures (non-discardable pictures) rather than all pictures. This partial application reduces processing complexity and avoids error-prone operations on discardable pictures that would not be used anyway, while still providing sufficient coverage for effective neural-network post-filtering.
4Adaptability or versatility
If NNPFA SEI messages are associated with non-output pictures, then filter activation flexibility is improved, but reliability deteriorates due to mismatch between encoder and decoder
Solution Approach 1:
The patent applies preliminary action by having the encoder signal the input picture list for the neural-network post-filter to the decoder before NNPFA messages are processed. This ensures that both encoder and decoder have consistent knowledge of which pictures should receive filter activation, preventing mismatches while maintaining the flexibility to activate the filter on appropriate non-output pictures.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are an image encoding/decoding method, a bitstream transmitting method, and a computer-readable recording medium for storing a bitstream. The image decoding method according to the present disclosure may include the steps of: acquiring unit information including a current picture; and decoding the current picture on the basis of the unit information, wherein a type of the current picture may be limited on the basis of a neural-network post-filter (NNPF)-related supplemental enhancement information (SEI) message included in the unit information.