Block Map Image Decoding with Adaptive Neighbor Availability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding and decoding methods struggle to efficiently handle high-resolution or high-quality image content by optimizing data units based on image characteristics, leading to suboptimal compression and decoding performance.
Innovation Solution
The method involves using a block map to determine decoded and non-decoded blocks, allowing for adaptive reference to neighboring regions during decoding and encoding, and updating the block map as blocks are processed, which can include square or non-square shapes and consider both spatial and temporal neighbors for motion information derivation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If image data is encoded using conventional codecs with fixed data units, then encoding simplicity is maintained, but compression efficiency for high-resolution images deteriorates
Solution Approach 1:
The image is divided into multiple blocks with different data unit sizes based on image characteristics. The encoding device segments the image into regions requiring different compression approaches, allowing high-resolution images to be processed with optimized data units for each region, thereby improving compression efficiency without uniformly increasing complexity across the entire image.
Solution Approach 2:
The data unit size is made dynamic rather than fixed, adapting to local image characteristics. The encoding device adjusts data unit sizes based on measured image properties in different regions, enabling flexible compression optimization that responds to actual image content rather than applying a uniform encoding scheme throughout.
2Manufacturing precision
If data units are uniformly sized for all image regions, then encoding process simplicity is maintained, but decoding accuracy for high-quality image content deteriorates
Solution Approach 1:
The block map is constructed in advance during the encoding process, storing information about data unit sizes and positions for each block before decoding occurs. This preliminary organization of block information enables the decoding device to accurately reconstruct the image with appropriate data unit sizes for each region without requiring complex real-time decisions during decoding.
Solution Approach 2:
The block map serves as an intermediary data structure that bridges the encoder and decoder, containing pre-computed information about block configurations. This intermediary allows the decoder to achieve high decoding accuracy by referencing the block map, which encapsulates the complex spatial and size relationships without requiring the decoder to perform complex calculations.
3Measurement precision
If neighboring regions are always referenced for motion information, then decoding speed is improved, but accuracy deteriorates when neighboring blocks are not yet decoded
Solution Approach 1:
The block map enables blocks to determine their own reference block availability based on pre-stored decoding order information. Each block can independently query the block map to identify which neighboring blocks are already decoded and can serve as references, eliminating the need for complex inter-block coordination or waiting mechanisms while ensuring accurate motion information derivation from actually decoded neighbors.
Data Source
AI summary
Provided are a method and device for encoding or decoding an image by using a block map. In the method and device for decoding an image according to an embodiment set forth herein, a bitstream of an encoded image is received, one or more blocks in a picture are determined using split information obtained from the bitstream, a block map indicating whether or not a previous block among the one or more blocks has been decoded is determined, a neighboring region which is available to be referred to for a current block to be decoded in a certain scan order among the one or more blocks is determined, and the current block is decoded based on the neighboring region which is available to be referred to.