Indexed Alpha Value Mapping for Compact Transparent Image Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image data formats that represent color depth with 10 bits per component and alpha value with 8 bits result in excessive storage requirements and limited transparency representation, leading to inefficient memory usage and longer access times.
Innovation Solution
An alpha value determination apparatus and method that uses a pixel data storage unit, alpha value data storage unit, and alpha value determination unit to manage alpha values through indices, allowing for reduced bit allocation while maintaining rich color and transparency representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If 10 bits are allocated for color depth and 8 bits for alpha value, then rich color and transparency representation is achieved, but storage capacity requirement doubles and memory access time increases
Solution Approach 1:
The patent segments the alpha value storage into two parts: a shared lookup table storing common alpha values and per-pixel index values pointing to these entries. This segmentation allows the system to maintain rich transparency representation while dramatically reducing per-pixel storage requirements from 8 bits to 2 bits.
Solution Approach 2:
The patent creates a shared lookup table that stores alpha values in common across multiple pixels. Instead of storing full 8-bit alpha values for each pixel, the system stores compact 2-bit indices that reference pre-defined alpha values in the lookup table, effectively copying the alpha value information in a compressed form.
2Loss of information
If 10 bits are allocated for color depth and 8 bits for alpha value, then rich color and transparency representation is achieved, but memory access time becomes longer
Solution Approach 1:
By segmenting alpha value storage into a shared lookup table and per-pixel indices, the system reduces the amount of data that needs to be accessed and transferred per pixel, thereby decreasing memory access time while preserving transparency quality.
Solution Approach 2:
The shared lookup table stores alpha values in a compressed, reusable format that can be quickly referenced by multiple pixels through small index values, reducing memory bandwidth requirements and access time compared to storing full alpha values for each pixel.
3Quantity of substance
If 2 bits are allocated for alpha value, then storage capacity is reduced, but transparency representation is limited to only two stages
Solution Approach 1:
The patent uses a shared lookup table that stores multiple distinct alpha values (e.g., 0x00, 0x80, 0xFF and intermediate values) that can be referenced by 2-bit indices. This allows the system to maintain compact per-pixel storage while recovering richer transparency representation through the shared data structure.
Solution Approach 2:
The patent transitions from storing alpha values directly in each pixel to storing references (indices) to a shared lookup table. This dimensional change allows the system to represent more transparency levels than the 2 bits would normally permit, by leveraging the shared storage dimension.
4Loss of information
If 9 bits are allocated for color depth and 5 bits for alpha value, then color and transparency representation is maintained, but storage efficiency is not optimized
Solution Approach 1:
The patent changes the parameter representation by introducing a shared lookup table structure. Instead of directly storing 5-bit or 8-bit alpha values per pixel, the system transforms the representation to use 2-bit indices referencing a shared table, optimizing storage efficiency while maintaining transparency quality.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Provided are an alpha value determination apparatus, an alpha value determination method, a program, and a data structure of image data that can represent an image rich in both color and transparency while suppressing the storage capacity necessary for storing the image data. A pixel data storage unit (32a) stores a plurality of pieces of pixel data indicating a pixel value and an index and associated individually with a plurality of pixels included in an image. An alpha value data storage unit (32b) stores alpha value data that is referenced in common by the index indicated in the plurality of pieces of pixel data and that indicates correspondence between the index and an alpha value. A combined image generation unit (38) determines alpha values of a plurality of pixels associated with the respective pieces of pixel data based on the index indicated in the pixel data and the alpha value data.