Character Recognition TRIZ Cases
- Optical Character Recognition Using Binary Tree Opcodes
Binary tree navigation using opcodes reduces processing time and improves accuracy for multiple languages.
- Image Legibility Assessment Using Text Component Analysis
A system evaluates image legibility by analyzing text components and applying trained classifiers to determine content usability.
- OCR False Recognition Correction via Character Association
An information processing apparatus associates recognized characters with specific accuracy indices to register and extract target strings.
- Augmented Image Correlation for Object Recognition
Extracts corners and edges from source images to augment traditional correlation, resolving recognition accuracy issues under varying lighting conditions.
- Adaptive Binarization Threshold Calculation via Local Histogram Statistics
A binarization threshold calculation system computes local histogram statistics to determine adaptive thresholds for grayscale images.
- Character String Region Detection in Image Processing
A detection module identifies character string regions within target object images using template matching algorithms.
- Client Side Card OCR Image Filtering
Local image analysis filters unusable card scans before transmission, reducing server processing capacity burden.
- OCR Text Graph Structuring for Overlapping Scenes
Text graph structuring converts overlapping text regions into topology graphs split into independent subgraphs for accurate character recognition.
- Ink Element Recognition via Perimeter-Based Stroke Construction
A recognition system determines ink element perimeters to construct strokes for accurate object identification.
- Gap Shifting for Tabular Text Recognition
Gap shifting reduces missing right-justified text by moving elements closer, improving OCR accuracy.
- Image Registration Using Character Position Matching
Character position matching replaces complex corner detection with labeled regions to improve registration accuracy and universality.
- Text Region Distortion Compensation via Affine Transform
Estimating boundary forms enables affine transforms to rectify geometric distortion in captured images, improving text recognition accuracy.
- Optical Character Recognition Text Input System
An optical character recognition module converts captured text images into digital characters, resolving input speed bottlenecks from manual keystrokes.
- Redaction Block Detection in Document OCR Systems
Segment document images to isolate blacked-out areas before optical character recognition, preventing garbage output from redacted regions.
- Video OCR Tracking via Common Coordinate System
A tracker system establishes a common coordinate system to align text recognition results across sequential video frames.
- Character Region Extraction via Stroke Width Calculation
A character region extracting apparatus calculates stroke widths along extracted outlines to determine character presence in candidate image regions.
- Batch Construction for Text Recognition via Width Thresholding
Sorts bounding boxes by width to construct batches with minimal padding, resolving the trade-off between processing speed and recognition accuracy.