Chinese Handwriting Ocr

PassAudited by VirusTotal on May 11, 2026.

Findings (1)

The bundle provides a comprehensive suite for Chinese handwriting OCR using PaddleOCR and RapidOCR engines. It includes scripts for synthetic data generation (scripts/generate_training_data.py), model training and fine-tuning (scripts/train_crnn.py, scripts/train_handwriting_model.py), and advanced inference pipelines with multi-pass voting and context-aware validation (scripts/ocr_smart_aligner.py, scripts/ocr_date_extractor.py). The code uses standard libraries like PyMuPDF (fitz), RapidOCR, and PyTorch for its stated purposes, and no malicious behaviors such as data exfiltration, unauthorized network access, or persistence mechanisms were detected.