360-Degree Image Encoding and Decoding Using Projection-Aware Block Division

The patent divides the 360-degree image into multiple projection formats (ERP, CMP, OHP, ISP) and processes each format separately with format-specific optimization. The decoding process is segmented into distinct stages: syntax information extraction, predicted image generation, residual image processing, and format-specific reconstruction. This segmentation allows the system to handle massive data volumes by processing different portions of the image data through specialized pipelines optimized for their respective projection characteristics.

The patent transforms the processing approach by introducing projection format as an additional dimension for organizing and processing 360-degree image data. Instead of treating all pixel data uniformly, the system reorganizes the massive data volume according to projection geometry (e.g., cube faces for CMP, spherical coordinates for ERP), enabling more efficient memory access patterns and parallel processing strategies that improve overall productivity despite the large data quantity.

The patent applies segmentation by dividing the image into partitioning units and processing each unit independently through expansion and rearrangement operations. Each partitioning unit is handled with format-specific rules, allowing the complex processing to be broken down into manageable, repeatable steps that can be parallelized, thereby improving compression performance without overwhelming the processing system.

The patent implements local quality by applying different expansion and rearrangement strategies to different regions of the image based on their projection format characteristics. For example, cube map projections may use face-based partitioning while equirectangular projections use latitude-longitude grid partitioning. This localized processing approach optimizes compression for each region's specific geometric properties while keeping the overall processing complexity manageable through consistent local rules.

The patent achieves universality by designing a single decoding apparatus that can handle multiple projection formats (ERP, CMP, OHP, ISP) through a unified architecture. The system uses format identification syntax information to route data through appropriate processing pipelines, allowing one device to serve multiple VR/AR application requirements without requiring separate specialized hardware for each projection type, thus improving adaptability while controlling complexity.

The patent applies preliminary action by extracting and identifying the projection format from syntax information at the beginning of the decoding process. This early format identification allows the system to pre-configure the appropriate processing pipeline and parameters before actual image reconstruction begins, enabling multi-format support without requiring complex real-time format switching or multiple complete decoding chains, thereby reducing overall system complexity.