Spaces:
Sleeping
Sleeping
File size: 1,693 Bytes
1a1c667 23fad62 1a1c667 3497db5 1a1c667 436a6c3 1a1c667 3497db5 706dc31 1a1c667 3497db5 1a1c667 436a6c3 5888c6a 706dc31 1471ff9 706dc31 5888c6a 1471ff9 1a1c667 1471ff9 8ceac8e 1471ff9 3497db5 1471ff9 986c935 1471ff9 5888c6a 1471ff9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 |
import os
import cv2
import numpy as np
from PIL import Image
import torch
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
# ✅ Load TrOCR model and processor once
try:
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-handwritten")
model.eval()
print("✅ TrOCR model loaded.")
trocr_available = True
except Exception as e:
print(f"❌ Failed to load TrOCR: {e}")
trocr_available = False
def preprocess_image(image):
"""
Preprocess image for OCR: convert to grayscale and enhance contrast.
"""
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
denoised = cv2.fastNlMeansDenoising(gray, h=10)
processed = cv2.adaptiveThreshold(
denoised, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY, 11, 2
)
return processed
def extract_text_from_image(image_path):
"""
Extract handwritten text from an image using TrOCR
"""
try:
print(f"📂 Reading image from: {image_path}") # ✅ log file path
image = Image.open(image_path).convert("RGB")
pixel_values = processor(images=image, return_tensors="pt").pixel_values
generated_ids = trocr_model.generate(pixel_values)
text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
print(f"📝 Extracted Text from {os.path.basename(image_path)}: {text}") # ✅ PRINT EXTRACTED TEXT
return text or "Text extraction failed."
except Exception as e:
print(f"❌ OCR failed on {image_path}: {str(e)}")
return "Text extraction failed."
|