Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -5,49 +5,60 @@ from PIL import Image
|
|
5 |
from paddleocr import PaddleOCR
|
6 |
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
7 |
import easyocr
|
|
|
|
|
8 |
|
9 |
-
# Initialize models
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
|
15 |
def run_paddleocr(image):
|
16 |
-
"""Run PaddleOCR on image"""
|
17 |
-
if isinstance(image, Image.Image):
|
18 |
-
image = np.array(image)
|
19 |
-
|
20 |
try:
|
21 |
-
result =
|
22 |
return ' '.join([line[1][0] for line in result[0]]) if result else ''
|
23 |
except Exception as e:
|
24 |
-
return f"
|
25 |
|
26 |
def run_easyocr(image):
|
27 |
-
"""Run EasyOCR on image"""
|
28 |
-
if isinstance(image, Image.Image):
|
29 |
-
image = np.array(image)
|
30 |
-
|
31 |
try:
|
32 |
-
result =
|
33 |
return ' '.join(result) if result else ''
|
34 |
except Exception as e:
|
35 |
-
return f"
|
36 |
|
37 |
def run_trocr(image):
|
38 |
-
"""Run TrOCR on image"""
|
39 |
-
if isinstance(image, np.ndarray):
|
40 |
-
image = Image.fromarray(image)
|
41 |
-
|
42 |
try:
|
43 |
-
pixel_values =
|
44 |
-
generated_ids =
|
45 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
except Exception as e:
|
47 |
-
return f"
|
48 |
|
49 |
def compare_models(image):
|
50 |
-
"""Compare all three OCR models"""
|
51 |
if isinstance(image, np.ndarray):
|
52 |
image = Image.fromarray(image)
|
53 |
image = image.convert("RGB")
|
@@ -55,75 +66,64 @@ def compare_models(image):
|
|
55 |
results = {}
|
56 |
times = {}
|
57 |
|
58 |
-
# Run
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
# Run TrOCR
|
69 |
-
start = time.time()
|
70 |
-
results['TrOCR'] = run_trocr(image)
|
71 |
-
times['TrOCR'] = time.time() - start
|
72 |
|
73 |
# Create comparison table
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
74 |
comparison = f"""
|
75 |
-
<table style="width:100%; border-collapse: collapse;">
|
76 |
<tr style="background-color: #f2f2f2;">
|
77 |
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">Model</th>
|
78 |
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">Extracted Text</th>
|
79 |
-
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">
|
80 |
-
</tr>
|
81 |
-
<tr>
|
82 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">PaddleOCR</td>
|
83 |
-
<td style="padding: 8px; border: 1px solid #ddd;">{results['PaddleOCR']}</td>
|
84 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">{times['PaddleOCR']:.3f}</td>
|
85 |
-
</tr>
|
86 |
-
<tr>
|
87 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">EasyOCR</td>
|
88 |
-
<td style="padding: 8px; border: 1px solid #ddd;">{results['EasyOCR']}</td>
|
89 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">{times['EasyOCR']:.3f}</td>
|
90 |
-
</tr>
|
91 |
-
<tr>
|
92 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">TrOCR</td>
|
93 |
-
<td style="padding: 8px; border: 1px solid #ddd;">{results['TrOCR']}</td>
|
94 |
-
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">{times['TrOCR']:.3f}</td>
|
95 |
</tr>
|
|
|
96 |
</table>
|
97 |
"""
|
98 |
|
99 |
-
return comparison, results
|
100 |
|
101 |
-
#
|
102 |
-
with gr.Blocks(title="
|
103 |
-
gr.Markdown(""
|
104 |
-
## English OCR Models Comparison
|
105 |
-
This tool compares three OCR models for English text:
|
106 |
-
1. PaddleOCR
|
107 |
-
2. EasyOCR
|
108 |
-
3. TrOCR (Microsoft)
|
109 |
-
""")
|
110 |
|
111 |
with gr.Row():
|
112 |
with gr.Column():
|
113 |
-
|
114 |
-
|
|
|
|
|
|
|
|
|
|
|
115 |
|
116 |
with gr.Column():
|
117 |
-
|
118 |
-
with gr.Accordion("
|
119 |
-
|
120 |
-
|
121 |
-
trocr_output = gr.Textbox(label="TrOCR Result")
|
122 |
|
123 |
submit_btn.click(
|
124 |
fn=compare_models,
|
125 |
-
inputs=
|
126 |
-
outputs=[
|
127 |
)
|
128 |
|
129 |
if __name__ == "__main__":
|
|
|
5 |
from paddleocr import PaddleOCR
|
6 |
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
7 |
import easyocr
|
8 |
+
import pytesseract
|
9 |
+
from doctr.models import ocr_predictor
|
10 |
|
11 |
+
# Initialize all models
|
12 |
+
models = {
|
13 |
+
"PaddleOCR": PaddleOCR(lang='en'),
|
14 |
+
"EasyOCR": easyocr.Reader(['en']),
|
15 |
+
"TrOCR": {
|
16 |
+
"processor": TrOCRProcessor.from_pretrained("microsoft/trocr-base-printed"),
|
17 |
+
"model": VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-printed")
|
18 |
+
},
|
19 |
+
"Tesseract": None, # Initialized by pytesseract
|
20 |
+
"DocTR": ocr_predictor(det_arch='db_resnet50', reco_arch='crnn_vgg16_bn', pretrained=True)
|
21 |
+
}
|
22 |
|
23 |
def run_paddleocr(image):
|
|
|
|
|
|
|
|
|
24 |
try:
|
25 |
+
result = models["PaddleOCR"].ocr(np.array(image))
|
26 |
return ' '.join([line[1][0] for line in result[0]]) if result else ''
|
27 |
except Exception as e:
|
28 |
+
return f"Error: {str(e)}"
|
29 |
|
30 |
def run_easyocr(image):
|
|
|
|
|
|
|
|
|
31 |
try:
|
32 |
+
result = models["EasyOCR"].readtext(np.array(image), detail=0)
|
33 |
return ' '.join(result) if result else ''
|
34 |
except Exception as e:
|
35 |
+
return f"Error: {str(e)}"
|
36 |
|
37 |
def run_trocr(image):
|
|
|
|
|
|
|
|
|
38 |
try:
|
39 |
+
pixel_values = models["TrOCR"]["processor"](image, return_tensors="pt").pixel_values
|
40 |
+
generated_ids = models["TrOCR"]["model"].generate(pixel_values)
|
41 |
+
return models["TrOCR"]["processor"].batch_decode(generated_ids, skip_special_tokens=True)[0]
|
42 |
+
except Exception as e:
|
43 |
+
return f"Error: {str(e)}"
|
44 |
+
|
45 |
+
def run_tesseract(image):
|
46 |
+
try:
|
47 |
+
return pytesseract.image_to_string(image, lang='eng')
|
48 |
+
except Exception as e:
|
49 |
+
return f"Error: {str(e)}"
|
50 |
+
|
51 |
+
def run_doctr(image):
|
52 |
+
try:
|
53 |
+
if isinstance(image, Image.Image):
|
54 |
+
image = np.array(image)
|
55 |
+
result = models["DocTR"]([image])
|
56 |
+
return ' '.join([word[0] for page in result.pages for block in page.blocks
|
57 |
+
for line in block.lines for word in line.words])
|
58 |
except Exception as e:
|
59 |
+
return f"Error: {str(e)}"
|
60 |
|
61 |
def compare_models(image):
|
|
|
62 |
if isinstance(image, np.ndarray):
|
63 |
image = Image.fromarray(image)
|
64 |
image = image.convert("RGB")
|
|
|
66 |
results = {}
|
67 |
times = {}
|
68 |
|
69 |
+
# Run all OCR models
|
70 |
+
for name, func in [("PaddleOCR", run_paddleocr),
|
71 |
+
("EasyOCR", run_easyocr),
|
72 |
+
("TrOCR", run_trocr),
|
73 |
+
("Tesseract", run_tesseract),
|
74 |
+
("DocTR", run_doctr)]:
|
75 |
+
start = time.time()
|
76 |
+
results[name] = func(image)
|
77 |
+
times[name] = time.time() - start
|
|
|
|
|
|
|
|
|
|
|
78 |
|
79 |
# Create comparison table
|
80 |
+
table_rows = []
|
81 |
+
for name in results:
|
82 |
+
table_rows.append(f"""
|
83 |
+
<tr>
|
84 |
+
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">{name}</td>
|
85 |
+
<td style="padding: 8px; border: 1px solid #ddd;">{results[name]}</td>
|
86 |
+
<td style="padding: 8px; border: 1px solid #ddd; text-align: center;">{times[name]:.3f}</td>
|
87 |
+
</tr>
|
88 |
+
""")
|
89 |
+
|
90 |
comparison = f"""
|
91 |
+
<table style="width:100%; border-collapse: collapse; margin-bottom: 20px;">
|
92 |
<tr style="background-color: #f2f2f2;">
|
93 |
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">Model</th>
|
94 |
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">Extracted Text</th>
|
95 |
+
<th style="padding: 8px; border: 1px solid #ddd; text-align: center;">Time (s)</th>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
96 |
</tr>
|
97 |
+
{''.join(table_rows)}
|
98 |
</table>
|
99 |
"""
|
100 |
|
101 |
+
return comparison, *results.values()
|
102 |
|
103 |
+
# Gradio Interface
|
104 |
+
with gr.Blocks(title="Advanced OCR Comparison") as demo:
|
105 |
+
gr.Markdown("## 🚀 Advanced English OCR Comparison (5 Models)")
|
|
|
|
|
|
|
|
|
|
|
|
|
106 |
|
107 |
with gr.Row():
|
108 |
with gr.Column():
|
109 |
+
img_input = gr.Image(label="Upload Document", type="pil")
|
110 |
+
gr.Examples(
|
111 |
+
examples=["sample1.jpg", "sample2.png"],
|
112 |
+
inputs=img_input,
|
113 |
+
label="Sample Images"
|
114 |
+
)
|
115 |
+
submit_btn = gr.Button("Run Comparison", variant="primary")
|
116 |
|
117 |
with gr.Column():
|
118 |
+
comparison = gr.HTML(label="Comparison Results")
|
119 |
+
with gr.Accordion("Detailed Results", open=False):
|
120 |
+
gr.Markdown("### Individual Model Outputs")
|
121 |
+
outputs = [gr.Textbox(label=name) for name in models]
|
|
|
122 |
|
123 |
submit_btn.click(
|
124 |
fn=compare_models,
|
125 |
+
inputs=img_input,
|
126 |
+
outputs=[comparison, *outputs]
|
127 |
)
|
128 |
|
129 |
if __name__ == "__main__":
|