surya_ocr_text_detection_and_recognition_quickstart.py

1from PIL import Image
2from surya.ocr import run_ocr
3from surya.model.detection.model import load_model as load_det_model, load_processor as load_det_processor
4from surya.model.recognition.model import load_model as load_rec_model
5from surya.model.recognition.processor import load_processor as load_rec_processor
6
7# Replace with the path to your image
8image_path = "input_image.png"
9image = Image.open(image_path)
10langs = ["en"] # Replace with your languages
11
12# Load models and processors
13det_processor = load_det_processor()
14det_model = load_det_model()
15
16rec_model = load_rec_model()
17rec_processor = load_rec_processor()
18
19# Run OCR
20# This will perform both text detection and recognition
21predictions = run_ocr([image], [langs], det_model, det_processor, rec_model, rec_processor)
22
23# Print results
24for result in predictions:
25    for line in result.text_lines:
26        print(f"Text: {line.text}")
27        print(f"Confidence: {line.confidence}")
28        print(f"Bbox: {line.bbox}")