Binary Volume Rendering via Bit Packing and Decoding Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern graphics processing units (GPUs) face challenges in efficiently storing and rendering binary volumes due to their inability to perform boolean operations for fragment processing or store texture formats for binary data, leading to increased memory requirements and loss of information when expanding binary segmentation masks into gray volumes.
Innovation Solution
A method and system for rendering binary volumes by storing eight bits along an x-axis in a byte, generating a two-dimensional decoding table, extracting and decoding bit coordinates, and multiplying decoded values by gray values to form resultant values, which are then blended into a frame buffer, reducing memory usage and preventing information loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If binary volumes are expanded by assigning a byte value to each bit, then the binary volume can be stored and rendered in GPU memory, but the storage requirements increase by a factor of 8, 16 or 32
Solution Approach 1:
The binary volume is segmented into bytes, with each byte containing 8 bits packed together. This segmentation allows the binary data to be stored in a compact format that is compatible with GPU memory requirements while maintaining the ability to render the volume efficiently.
Solution Approach 2:
The patent introduces a decoding table that maps from the packed byte representation to the original bit values. This dimensional transformation allows the system to work with compact byte data while preserving the ability to access individual bit values for rendering operations.
2Quantity of substance
If a binary segmentation mask is burned into a gray volume, then memory requirements are reduced, but information is lost and artifacts appear in rendered images
Solution Approach 1:
Instead of burning the binary mask into the gray volume (which loses information), the patent extracts the binary mask data into a separate compact representation. This allows the original gray volume data to be preserved while the binary segmentation information is stored separately in a space-efficient format.
Solution Approach 2:
The decoding table serves as an intermediary that bridges the compact byte representation and the original bit values. This intermediary structure allows the system to reduce memory usage while maintaining full information distinguishability through the lookup table mechanism.
3Quantity of substance
If binary volumes are stored in their native form with bits packed in bytes, then memory requirements are reduced, but modern GPUs cannot perform boolean operations or store texture formats for binary data
Solution Approach 1:
The patent changes the data representation parameters by packing 8 bits into bytes and creating a decoding table that translates between the packed format and standard GPU texture formats. This parameter transformation enables compatibility with modern GPU architectures while maintaining compact storage.
Data Source
AI summary
A system and method for rendering a binary volume in a GPU are provided. The method for rendering a binary volume comprises: storing eight bits of the binary volume along an x-axis in a byte of memory to form a compact representation of the binary volume; generating a two-dimensional decoding table of the compact representation of the binary volume; extracting a byte including a bit of interest from the compact representation of the binary volume; determining a bit coordinate for the bit of interest and extracting a decoded value of the bit of interest from the two-dimensional decoding table; obtaining a gray value of the decoded value of the bit of interest and classifying the gray value; and multiplying the decoded value of the bit of interest by the gray value to form a resultant value.