Character Recognition TRIZ Cases
- Log Linear Model for OCR Feature Weight Optimization
A log linear model evaluates feature functions to determine transcription costs for text images.
- Portable Scanner with Local Storage for Offline Data Capture
Handheld optical scanner integrates local memory to capture text and images, eliminating the need for constant wired connections during data acquisition.
- Removing Connective Strokes in Cursive Handwriting Recognition
A processing system identifies connective points using connective scores to remove strokes from contiguous handwritten input.
- Document Image Cleansing via Adaptive Cleaning and OCR Selection
Adaptive cleaning removes deletion targets while dynamic OCR selection resolves the trade-off between processing complexity and character recognition accuracy.
- Optical Character Recognition Text Extraction for Mobile Input
Optical character recognition algorithms convert captured text images into digital strings, eliminating manual typing delays on portable devices.
- Post-OCR Shape Clustering for Character Code Correction
A post-OCR processing system classifies clip images into shape clusters to generate representative images for character code identification.
- Raster Image Text Detection via Shape Chain Straightening
A raster image text detection system converts shapes into vector chains and straightens curved segments for classification.
- Dynamic OCR System with User Feedback Loop
A dynamic optical character recognition system creates customized machine learning algorithms based on user input to expand its character set.