| import gradio as gr | |
| from craft_hw_ocr import OCR | |
| ocr = OCR.load_models() | |
| def do_ocr(inp): | |
| img, results = OCR.detection(inp, ocr[2]) | |
| bboxes, text = OCR.recoginition(img, results, ocr[0], ocr[1]) | |
| return OCR.visualize(img, results), text | |
| inputs = gr.inputs.Image() | |
| o1 = gr.outputs.Image() | |
| o2 = gr.outputs.Textbox() | |
| title = "CRAFT-OCR" | |
| description = "OCR of both handwriting and printed text using CRAFT Text detector and TrOCR recognition, detection of lines and extraction of them are happening here because TrOCR pre-trained models are modelled on IAM lines dataset and the same needs to be implemented here." | |
| examples=[['example_1.png'],['example_2.jpg']] | |
| article = "<p style='text-align: center'><a href='https://github.com/Vishnunkumar/craft_hw_ocr' target='_blank'>craft_hw_ocr</a></p><p style='text-align: center'> <p style='text-align: center'><a href='https://github.com/fcakyon/craft-text-detector' target='_blank'>craft-text-detector</a></p><p style='text-align: center'>" | |
| gr.Interface(fn=do_ocr, inputs=inputs, outputs=[o1, o2], title=title, description=description, article=article, examples=examples, enable_queue=True).launch() | |