PP OCRv5 Mobile Det

PP-OCRv5_mobile_det is a textline detection model that forms part of the PP-OCRv5 series released by PaddlePaddle. Hosted on Hugging Face, it addresses the need for accurate detection of text regions in images as the first stage of an optical character recognition pipeline. The model carries the apache-2.0 license and is provided for direct use in code or notebooks.

It belongs to the PP-OCRv5_det family and supports English and Chinese text. Integration occurs through the PaddleOCR library after installing the paddlepaddle backend and the paddleocr Python package. Example code loads the detector with the line TextDetection(model_name="PP-OCRv5_mobile_det"), runs inference on an image path, and outputs results that can be printed, saved as images, or exported to JSON.

The repository supplies notebooks for Google Colab and Kaggle to facilitate experimentation. Documentation links cover installation, quick start, model usage, and pipeline integration. As an image-to-text component focused on textline_detection, it is listed alongside references to related arXiv work and carries tags for PaddleOCR, OCR, and the supported languages.

PaddlePaddle maintains the model card, which includes files and version history. The entry indicates 28 likes and sits within the organization's collection of 3.39k models.