Text Encoding with Bitmap Indexing for Fast Longest-Match Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing longest coincidence searching method using a slide window in ZIP compression algorithms faces challenges in improving compression velocity and compression ratio due to limited window size and sequential data comparison, which results in decreased performance and efficiency.
Innovation Solution
The proposed solution involves using a bit map type index to encode character strings by associating appearance positions with word IDs, allowing for efficient longest coincidence searching without the limitations of a slide window, and registering coincident strings in a dynamic dictionary for dynamic coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If longest coincidence searching is performed using a slide window in ZIP compression, then compression is achieved by representing codes with relative addresses and lengths, but the compression velocity decreases due to sequential data comparison and the compression ratio decreases due to limited window size
Solution Approach 1:
The patent replaces the mechanical sequential comparison process with a bitmap-based parallel search mechanism. Instead of sequentially comparing characters in a slide window, the invention uses pre-generated bitmaps where each bit represents the presence of a character at a specific position. This allows parallel processing of multiple character positions simultaneously, dramatically improving compression velocity while reducing searching time.
Solution Approach 2:
The invention performs preliminary actions by pre-generating bitmaps that store the positions of all characters in the compressed data before the actual compression search begins. These bitmaps are created in advance and stored in memory, allowing the compression process to quickly query character positions without performing sequential comparisons during compression, thus improving both velocity and reducing time loss.
2Productivity
If the slide window size is limited to prevent excessively long searching time and code length, then searching time is controlled, but the compression ratio decreases because character strings overflowing from the window are excluded from longest coincidence targets
Solution Approach 1:
The patent transitions from a one-dimensional slide window approach to a two-dimensional bitmap structure where one dimension represents character positions and the other represents character types. This dimensional change allows the system to access character position information across the entire compressed data set without being constrained by a fixed window size, enabling longer coincidence searches while maintaining efficient access times.
Solution Approach 2:
The invention creates a bitmap copy of the compressed data that stores character position information in a condensed format. Instead of storing and searching through the actual character data in a large window, the system uses this bitmap copy to quickly determine character positions and perform coincidence searches, effectively removing the window size limitation while maintaining searching efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An encoding apparatus (100) reads text data of an encoding target, encodes each character or word in the text data of the encoding target by using a bit map type index (123) in which an appearance position is associated with each of the encoded characters or words, appearing on the text data of the encoding target, as bit map data, and updates the bit map type index with respect to the encoded character or word.