Handwriting Retrieval via Sub-Sequence Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data retrieval methods from databases using point sequences are inefficient in handling handwritten queries, particularly in accurately matching or retrieving similar sequences due to limitations in processing and analyzing the complexity of handwritten data.
Innovation Solution
An apparatus comprising an acquirer, divider, extractor, calculator, and retriever that acquires and processes handwritten point sequences by dividing them into sub-sequences, extracting feature vectors, calculating distances using dynamic programming, and retrieving data based on optimal path matching, thereby enhancing retrieval precision and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If handwritten point sequences are processed as complete sequences, then retrieval accuracy may be maintained, but processing complexity and time increase significantly
Solution Approach 1:
The patent divides handwritten point sequences into multiple sub-sequences based on spatial relationships and stroke characteristics. This segmentation allows the system to process smaller, more manageable units rather than treating the entire sequence as one complex object, thereby reducing processing time while maintaining retrieval accuracy through hierarchical matching
2Measurement precision
If dynamic programming is applied to match entire point sequences, then optimal path matching is achieved, but computational cost increases
Solution Approach 1:
The patent applies dynamic programming to matched sub-sequences rather than entire sequences. By segmenting the point sequences first and then applying DP matching only to corresponding sub-sequences, the computational complexity is dramatically reduced while still achieving optimal path matching within each segment, which collectively provides accurate overall matching
Solution Approach 2:
The patent extracts feature vectors from sub-sequences to represent their essential characteristics. This extraction process identifies key features that capture the essence of each sub-sequence, allowing the system to perform matching based on these condensed representations rather than processing all raw point data, thereby reducing computational complexity
3Reliability
If feature extraction is performed on all point sequences, then comprehensive data is obtained, but processing overhead increases
Solution Approach 1:
The patent performs feature extraction on segmented sub-sequences rather than on complete point sequences. This segmentation approach maintains data completeness by ensuring features are extracted from all relevant portions of the handwriting, while improving processing efficiency by reducing the size and complexity of the data that needs to be processed at once
Data Source
AI summary
According to an embodiment, an apparatus includes an acquirer, a divider, an extractor, a calculator, and a retriever. The acquirer is configured to acquire a plurality of first point sequences. The divider is configured to divide each of the plurality of first point sequences into a plurality of second point sequences. The extractor is configured to extract a feature vector of each of the plurality of second point sequences. The calculator is configured to calculate a distance between the plurality of first point sequences based on an optimal path among the plurality of second point sequences that belong to each of the plurality of first point sequences. The retriever is configured to retrieve data corresponding to the plurality of first point sequences using the distance.