Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,45 +1,59 @@
|
|
| 1 |
import cv2
|
| 2 |
import pytesseract
|
| 3 |
from pytesseract import Output
|
| 4 |
-
import gradio as gr
|
| 5 |
import numpy as np
|
|
|
|
| 6 |
from PIL import Image
|
|
|
|
|
|
|
| 7 |
|
| 8 |
-
def
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
# ลด noise
|
| 13 |
-
gray = cv2.medianBlur(gray, 3)
|
| 14 |
|
| 15 |
-
#
|
| 16 |
-
|
| 17 |
-
gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
|
| 18 |
-
cv2.THRESH_BINARY, 11, 2
|
| 19 |
-
)
|
| 20 |
-
return thresh
|
| 21 |
-
|
| 22 |
-
def ocr_image(img_pil):
|
| 23 |
-
# แปลง PIL -> OpenCV
|
| 24 |
-
img = cv2.cvtColor(np.array(img_pil), cv2.COLOR_RGB2BGR)
|
| 25 |
|
| 26 |
-
#
|
| 27 |
-
|
| 28 |
|
| 29 |
-
|
| 30 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
title="OCR ไทย + อังกฤษ (Preprocessed)",
|
| 42 |
-
description="อัปโหลดรูปเพื่ออ่านข้อความภาษาไทยและอังกฤษ พร้อมปรับภาพให้แม่นยำขึ้น"
|
| 43 |
)
|
| 44 |
|
| 45 |
-
|
|
|
|
|
|
| 1 |
import cv2
|
| 2 |
import pytesseract
|
| 3 |
from pytesseract import Output
|
|
|
|
| 4 |
import numpy as np
|
| 5 |
+
import gradio as gr
|
| 6 |
from PIL import Image
|
| 7 |
+
from collections import defaultdict
|
| 8 |
+
import re
|
| 9 |
|
| 10 |
+
def ocr_with_boxes(image):
|
| 11 |
+
# แปลงจาก PIL → OpenCV
|
| 12 |
+
img = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
+
# config ไทย + อังกฤษ
|
| 15 |
+
custom_config = r'-l tha+eng --oem 3 --psm 6'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
+
# OCR with bounding boxes
|
| 18 |
+
data = pytesseract.image_to_data(img, config=custom_config, output_type=Output.DICT)
|
| 19 |
|
| 20 |
+
totalBox = len(data['text'])
|
| 21 |
+
lines = defaultdict(list)
|
| 22 |
+
|
| 23 |
+
for i in range(totalBox):
|
| 24 |
+
if int(data['conf'][i]) > 0:
|
| 25 |
+
text = data['text'][i].strip()
|
| 26 |
+
if text != "":
|
| 27 |
+
line_id = (data['block_num'][i], data['par_num'][i], data['line_num'][i])
|
| 28 |
+
lines[line_id] += [text]
|
| 29 |
+
|
| 30 |
+
# วาดกรอบ
|
| 31 |
+
x, y, w, h = data['left'][i], data['top'][i], data['width'][i], data['height'][i]
|
| 32 |
+
cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 2)
|
| 33 |
+
cv2.putText(img, text, (x, y - 5), cv2.FONT_HERSHEY_SIMPLEX,
|
| 34 |
+
0.7, (0, 0, 255), 2, cv2.LINE_AA)
|
| 35 |
+
|
| 36 |
+
# รวมข้อความให้เป็นบรรทัด
|
| 37 |
+
extracted_texts = [" ".join(words) for _, words in sorted(lines.items())]
|
| 38 |
+
final_text = "\n".join(extracted_texts)
|
| 39 |
+
|
| 40 |
+
# 🔧 แก้เว้นวรรคภาษาไทยตรงนี้เลย
|
| 41 |
+
final_text = re.sub(r'([\u0E00-\u0E7F])\s+([\u0E00-\u0E7F])', r'\1\2', final_text)
|
| 42 |
+
|
| 43 |
+
# แปลงกลับเป็น PIL
|
| 44 |
+
img_out = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
|
| 45 |
+
img_pil = Image.fromarray(img_out)
|
| 46 |
|
| 47 |
+
return img_pil, final_text
|
| 48 |
+
|
| 49 |
+
# Gradio UI
|
| 50 |
+
demo = gr.Interface(
|
| 51 |
+
fn=ocr_with_boxes,
|
| 52 |
+
inputs=gr.Image(type="pil", label="อัปโหลดภาพ"),
|
| 53 |
+
outputs=[gr.Image(label="ผลลัพธ์พร้อมกรอบข้อความ"), gr.Textbox(label="ข้อความ OCR")],
|
| 54 |
+
title="OCR ไทย + อังกฤษ (Tesseract + Fix Spacing)",
|
| 55 |
+
description="อัปโหลดภาพ ระบบจะ OCR ไทย+อังกฤษ วาดกรอบ และแก้เว้นวรรคภาษาไทยอัตโนมัติ"
|
|
|
|
|
|
|
| 56 |
)
|
| 57 |
|
| 58 |
+
if __name__ == "__main__":
|
| 59 |
+
demo.launch()
|