---
license: cc-by-4.0
pretty_name: 'Korean Receipts Dataset'
language:
- ko
tags:
- image
- korean
- receipts
- retail
- text-recognition
- ocr
- computer-vision
- ai-research
- multilingual
- finance
task_categories:
- image-classification
- text-recognition
- document-understanding
size_categories:
- 1K<n<10K
---

# Korean Receipts Dataset
*This dataset contains high-resolution images of Korean retail receipts from supermarkets, restaurants, and stores. The dataset has been anonymized to remove personal information and is intended for AI research in OCR, document understanding, and financial analytics.*

## Contact
For queries or collaborations related to this dataset, contact:  
  - anoushka@kgen.io  
  - abhishek.vadapalli@kgen.io  

## Supported Tasks

- **Task Categories**:  
  - Image Classification  
  - Text Recognition (OCR)  
  - Document Understanding  

- **Supported Tasks**:  
  - Text extraction from Korean receipts (item names, prices, totals, dates)  
  - Receipt classification by store type or transaction type  
  - Multilingual OCR for bilingual receipts (Korean-English)  
  - Financial data parsing and analytics  
  - Research in automated expense tracking and accounting systems  

## Languages

- **Primary Language**: Korean  
- **Secondary Presence**: English (on receipts from international or chain stores)  

## Dataset Creation

### Curation Rationale
This dataset was created to help AI systems accurately extract structured financial information from receipts. It supports OCR, text parsing, and document understanding models in retail and finance applications.

### Source Data
- **Contributors**: Field collection, simulated receipts, and crowdsourced images  
- **Collection Process**: Receipts were photographed or scanned. All personal identifiers (names, addresses, payment info) were removed or anonymized before inclusion.  

### Other Known Limitations
- **Bias**: Urban and chain store receipts may dominate; small vendor receipts are less represented  
- **Image Quality Variability**: Lighting, folds, or low-resolution printing may affect OCR accuracy  
- **Receipt Format Diversity**: Variation in layouts, fonts, and lengths may pose challenges for text extraction  

## Intended Uses

### ✅ Direct Use
- Training OCR models for Korean receipts  
- Document understanding and financial data extraction  
- Automated expense tracking applications  
- Academic research in computer vision and structured text parsing  

### ❌ Out-of-Scope Use
- Reconstructing personal transaction history  
- Commercial misuse of store or brand data without consent  
- Using receipts for surveillance or tracking individuals  

## License

CC BY 4.0