Handwritten Pattern Recognition via Core Point Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for recognizing cursive handwriting are complex and error-prone due to the difficulty in segmenting connected characters, requiring large training sets and significant processing power, and are highly dependent on dictionaries.
Innovation Solution
A method that quickly segments and recognizes handwritten patterns by selecting core points along the curves, determining features, and comparing them to templates, eliminating the need for neural networks and hidden Markov models, and allowing for efficient recognition of cursive handwriting without extensive learning or high processing power.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complicated statistical systems using neural networks and hidden markov models are employed to recognize cursive handwriting, then recognition accuracy is improved, but device complexity and processing requirements increase significantly
Solution Approach 1:
The handwritten pattern is segmented into multiple possible segmentations, each representing different ways to divide the continuous strokes into separate characters. The system generates a set of candidate segmentations and evaluates each one, rather than relying on a single complex statistical model to determine the correct segmentation automatically.
Solution Approach 2:
The system performs recognition for all possible segmentations (excessive action) rather than attempting to identify only the correct segmentation. By evaluating multiple segmentations and selecting the best match, the system achieves accurate recognition without requiring complex probabilistic models to predict the correct segmentation in advance.
2Reliability
If neural networks and hidden markov models with integrated dictionaries are used, then recognition of cursive handwriting is improved, but large training sets and significant processing power are required
Solution Approach 1:
The recognition process is segmented into distinct steps: generating possible segmentations, extracting features for each segmentation, comparing against templates, and selecting the best match. This stepwise approach allows for efficient processing by breaking down the complex recognition task into manageable operations that can be performed with reduced computational resources.
Solution Approach 2:
The system uses template matching where pre-defined character templates are compared against the segmented handwritten patterns. Instead of training complex neural networks, the system copies and compares against a library of standard character templates, significantly reducing the need for extensive training data and processing power.
3Reliability
If the handwritten pattern is segmented into multiple possible segmentations for recognition, then recognition accuracy is improved, but the number of comparisons to templates increases
Solution Approach 1:
The system applies different processing strategies to different parts of the recognition process. For each possible segmentation, only the relevant local features and comparisons are performed, rather than processing the entire pattern uniformly. This allows for efficient evaluation of multiple segmentations by focusing computational effort on local characteristics of each candidate segmentation.
Solution Approach 2:
The system performs preliminary processing by pre-defining character templates and pre-establishing comparison criteria before actual recognition occurs. This preliminary preparation allows for faster real-time recognition by eliminating the need to compute complex comparisons during the actual recognition process, thus reducing processing time despite evaluating multiple segmentations.
Data Source
AI summary
A method for recognition of a handwritten pattern comprises selecting core points among a sequence of detected points of the handwritten pattern. The core points are selected for use in segmenting and recognizing the handwritten pattern. The method may further comprise identifying potential diacritics in the sequence of points of the handwritten pattern, determining features of each core point, assigning at least one feature associated with identified potential diacritics to each core point of a subset of core points, and comparing the handwritten pattern to templates.