Remote Depth Buffer Compression via Tile Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for transmitting high dynamic range (HDR) depth images in remote rendering systems introduce discontinuities and significant latency due to inefficient processing of HDR depth images into low dynamic range (LDR) images, leading to spurious high frequency components and increased network bandwidth requirements.
Innovation Solution
The method involves decomposing HDR depth buffers into tiles, determining piecewise bilinear bounding functions iteratively to minimize quantization errors, and encoding LDR depth buffers with compressed tile data for transmission, using video encoders and lossless compressors, which are then reconstructed on the head-mounted device (HMD) with minimal computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If HDR depth images are converted to LDR images using conventional methods, then the dynamic range is reduced for transmission, but discontinuities and spurious high frequency components are introduced
Solution Approach 1:
The patent divides the depth image into multiple tiles and processes each tile independently with piecewise bilinear bounding functions. This segmentation allows localized optimization that preserves depth accuracy while enabling efficient LDR conversion and compression for transmission.
Solution Approach 2:
The patent transforms the depth image representation by changing parameters through piecewise bilinear bounding functions that map HDR depth values to LDR space. This parameter transformation maintains depth relationships and reduces quantization errors while achieving compact representation for transmission.
2Measurement precision
If HDR depth images are transmitted directly, then depth accuracy is maintained, but network bandwidth requirements increase significantly
Solution Approach 1:
The patent extracts only the essential depth information by fitting piecewise bilinear bounding functions to tile regions. Instead of transmitting full HDR depth images, only the bounding function parameters are transmitted, dramatically reducing bandwidth requirements while preserving depth accuracy for reprojection.
Solution Approach 2:
The patent changes the representation parameters from full HDR depth pixel values to compact LDR bounding function parameters. This parameter transformation enables efficient compression and transmission while maintaining the ability to reconstruct accurate depth information at the receiving end.
3Loss of time
If conventional LDR conversion methods are used, then transmission latency is reduced, but quantization errors increase
Solution Approach 1:
The patent segments the depth buffer into tiles and applies piecewise bilinear bounding functions to each tile. This localized approach reduces quantization errors within each tile while maintaining fast processing suitable for real-time transmission with minimal latency.
Solution Approach 2:
The patent uses iterative optimization of bounding function parameters based on feedback from depth buffer analysis. This feedback mechanism refines the LDR conversion to minimize quantization errors while maintaining efficient processing for low-latency transmission.
Data Source
AI summary
A high dynamic range (HDR) depth buffer is received at a remote computer. The HDR depth buffer is formed into a plurality of tiles. For each tile, a respective maximum and minimum value of the HDR depth buffer in a region greater than a width of the respective tile is determined. An initial pair of piecewise bilinear bounding functions for the HDR depth buffer is determined using the determined maximum and minimum depth values. For each tile, the initial pair of piecewise bilinear bounding functions is iteratively adjusted to move the respective minimum and maximum depth value of each tile closer to the HDR depth buffer, wherein no adjacent tile is adjusted in the same iteration. Using the adjusted pair of piecewise bilinear bounding functions, a low dynamic range (LDR) depth buffer and tile data are generated and encoded using a video encoder and a lossless compressor respectively.