matlab/matlab-agentic-toolkit / skills-catalog/image-processing-and-computer-vision/matlab-recognize-text
matlab-recognize-text
VerifiedBuild OCR pipelines in MATLAB using the ocr() function. Use this skill when the user wants to read text from images, documents, signs, meters, displays, license plates, gauges, receipts, or seven-segment displays. Covers image preprocessing, text detection (CRAFT, MSER), ROI-based recognition, multi-language OCR, and custom model training. Use when: OCR, text recognition, extract text from image, character recognition, document scanning, meter reading, gauge reading, receipt scanning, digitize text from photo.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 20.5 KB |
| manifest.yaml | config | 485 B |
| reference/batch-pipeline.md | markdown | 2.3 KB |
| reference/detection-methods.md | markdown | 3.6 KB |
| reference/function-reference.md | markdown | 3.8 KB |
| reference/limitations.md | markdown | 2.4 KB |
| reference/preprocessing-guide.md | markdown | 8.1 KB |
| reference/supported-languages.md | markdown | 2.9 KB |
| reference/training-guide.md | markdown | 5.6 KB |
reference/limitations.md
View on GitHub# OCR Limitations
**Tell the user early** when their image falls outside what MATLAB OCR can handle. Do not waste their time with endless preprocessing attempts on fundamentally unsolvable cases.
## Unsolvable Cases
| Image Type | Why OCR Cannot Work | What to Tell the User |
|---|---|---|
| Cursive handwriting | The OCR engine is trained on printed characters; connected script is unrecognizable regardless of preprocessing | "MATLAB's ocr() cannot read cursive handwriting. You would need a dedicated handwritten text recognition (HTR) model trained on handwriting datasets." |
| Artistic/brush calligraphy, WordArt | Stylized letterforms bear no resemblance to training data | "These decorative fonts are outside the OCR engine's capability. Consider trainOCR with labeled examples of this specific style." |
| CAPTCHAs | Intentionally designed to defeat automated text recognition | "CAPTCHAs are engineered to resist OCR. Expect <50% character accuracy even with heavy preprocessing." |
| Extremely degraded images (<30px text height) | Insufficient pixel information to distinguish characters even after upscaling | "The text is too small to recover — upscaling beyond 8-10x introduces more artifacts than information. A higher-resolution source image is needed." |
| Curved/arc text (badges, seals) | No single rotation straightens all characters; per-character segmentation is fragile | "Text on a curved path requires specialized dewarping. The skill's Hough deskewing handles linear skew, not arcs. Consider per-character detection or a custom approach." |
## How to Communicate Limitations
When you hit a limitation:
1. Inform the user clearly in 1-2 sentences
2. Explain *why* it won't work (not just that it won't)
3. Suggest the next step (trainOCR, higher-res source, different tool)
Do not silently loop through preprocessing attempts on a case you can diagnose as unsolvable from the initial inspection.
## Boundary Cases (Might Work with Effort)
| Image Type | Approach | Expected Outcome |
|---|---|---|
| Heavily degraded printed text | SAM + resize + CharacterSet constraint | ~60-80% accuracy possible |
| Non-Latin scripts (CJK, Arabic) | Built-in language models | Good for printed; fails on handwritten |
| Low-contrast embossed text | SAM segmentation | Works if texture provides contrast |
| Partially occluded text | Crop visible portions, OCR separately | Partial results only |
----
Copyright 2026 The MathWorks, Inc.
----
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |