Incremental Character Recognition via Pixel Segmentation and Confidence Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches for character recognition in images require significant computing resources, time, and can be inaccurate, making them inefficient and inconvenient for media processing tasks.
Innovation Solution
The system identifies a set of pixels depicting characters, determines linear portions within the image, and uses confidence scores from a language model to recognize characters, combining first and second confidence scores to meet a specified threshold for accurate character recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional character recognition approaches are used, then character recognition can be performed, but significant computing resources and time are required
Solution Approach 1:
The patent segments the character recognition process into distinct stages: pixel identification, linear portion detection, sub-portion analysis, and confidence scoring. This segmentation allows the system to process images incrementally, focusing computational resources only on relevant regions rather than analyzing the entire image, thereby improving processing efficiency while maintaining recognition accuracy
Solution Approach 2:
The patent performs preliminary actions by first identifying pixels that likely depict characters, then narrowing down to linear portions, and finally analyzing sub-portions. This preliminary filtering and progressive refinement approach reduces the computational burden by eliminating non-character regions before detailed analysis, thus improving productivity without sacrificing reliability
2Productivity
If conventional character recognition approaches are used, then character recognition can be performed, but the process is inaccurate
Solution Approach 1:
The patent implements feedback mechanisms through confidence scoring at multiple stages. The system generates confidence scores for pixels, linear portions, and sub-portions, using this feedback to iteratively refine the recognition process. This feedback-driven approach ensures that only high-confidence regions proceed to detailed analysis, improving measurement precision while maintaining productive processing speeds
Solution Approach 2:
The patent employs dynamic thresholding and adaptive processing where the analysis depth adjusts based on confidence scores. Regions with high confidence scores receive streamlined processing, while ambiguous regions trigger more detailed analysis. This dynamic approach optimizes the balance between processing speed and recognition accuracy by allocating computational resources adaptively
3Reliability
If conventional character recognition approaches are used, then character recognition can be performed, but significant computing resources are required
Solution Approach 1:
The patent extracts and focuses computational effort only on regions containing characters by identifying specific pixels, linear portions, and sub-portions that are likely to contain character data. This extraction approach eliminates the need to process entire images or irrelevant regions, significantly reducing computing resource consumption while maintaining recognition accuracy through targeted analysis of extracted character-containing regions
Data Source
AI summary
Systems, methods, and non-transitory computer-readable media can acquire an image that depicts at least one character. A set of pixels, within the image, through which the at least one character is depicted can be identified. At least one linear portion, within the image, can be identified based on the set of pixels. For each sub-portion within the at least one linear portion, a respective first confidence score representing a respective first likelihood that a respective sub-portion depicts the at least one character can be determined.