Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,26 +1,23 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
from PIL import Image
|
| 3 |
import easyocr
|
| 4 |
-
import cv2
|
| 5 |
import numpy as np
|
| 6 |
-
import
|
| 7 |
import pytesseract
|
| 8 |
|
| 9 |
# Initialize OCR models
|
| 10 |
easyocr_reader = easyocr.Reader(['en'])
|
| 11 |
-
paddleocr_reader =
|
| 12 |
-
# Tesseract
|
| 13 |
|
| 14 |
-
# Simulation of metrics, as per user scenario
|
| 15 |
def get_metrics(model):
|
| 16 |
-
# Simulated fixed scores according to user
|
| 17 |
-
base = 0.95
|
| 18 |
if model == "EasyOCR":
|
| 19 |
-
return {"Accuracy":
|
| 20 |
elif model == "PaddleOCR":
|
| 21 |
-
return {"Accuracy":
|
| 22 |
elif model == "Tesseract":
|
| 23 |
-
return {"Accuracy":
|
| 24 |
else:
|
| 25 |
return {}
|
| 26 |
|
|
@@ -31,11 +28,11 @@ def ocr_all(img, ground_truth):
|
|
| 31 |
metrics_easy = get_metrics("EasyOCR")
|
| 32 |
|
| 33 |
# PaddleOCR
|
| 34 |
-
result_paddle
|
| 35 |
try:
|
| 36 |
text_paddle = " ".join([line[1][0] for line in result_paddle])
|
| 37 |
except Exception:
|
| 38 |
-
text_paddle = "Error"
|
| 39 |
metrics_paddle = get_metrics("PaddleOCR")
|
| 40 |
|
| 41 |
# Tesseract
|
|
@@ -61,4 +58,4 @@ with gr.Blocks() as demo:
|
|
| 61 |
|
| 62 |
btn.click(ocr_all, inputs=[img_in, txt_in], outputs=[out_easy, out_paddle, out_tess])
|
| 63 |
|
| 64 |
-
demo.launch()
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
from PIL import Image
|
| 3 |
import easyocr
|
|
|
|
| 4 |
import numpy as np
|
| 5 |
+
from paddleocr import PaddleOCR
|
| 6 |
import pytesseract
|
| 7 |
|
| 8 |
# Initialize OCR models
|
| 9 |
easyocr_reader = easyocr.Reader(['en'])
|
| 10 |
+
paddleocr_reader = PaddleOCR()
|
| 11 |
+
# Tesseract uses pytesseract directly
|
| 12 |
|
|
|
|
| 13 |
def get_metrics(model):
|
| 14 |
+
# Simulated fixed scores according to user specs
|
|
|
|
| 15 |
if model == "EasyOCR":
|
| 16 |
+
return {"Accuracy": "95%", "Precision": "94%", "Pipeline": "Easy Integration (90%)"}
|
| 17 |
elif model == "PaddleOCR":
|
| 18 |
+
return {"Accuracy": "90%", "Precision": "89%", "Pipeline": "Moderate Integration (85%)"}
|
| 19 |
elif model == "Tesseract":
|
| 20 |
+
return {"Accuracy": "85%", "Precision": "83%", "Pipeline": "Manual Integration (75%)"}
|
| 21 |
else:
|
| 22 |
return {}
|
| 23 |
|
|
|
|
| 28 |
metrics_easy = get_metrics("EasyOCR")
|
| 29 |
|
| 30 |
# PaddleOCR
|
| 31 |
+
result_paddle = paddleocr_reader.ocr(np.array(img), cls=True)
|
| 32 |
try:
|
| 33 |
text_paddle = " ".join([line[1][0] for line in result_paddle])
|
| 34 |
except Exception:
|
| 35 |
+
text_paddle = "Error during PaddleOCR"
|
| 36 |
metrics_paddle = get_metrics("PaddleOCR")
|
| 37 |
|
| 38 |
# Tesseract
|
|
|
|
| 58 |
|
| 59 |
btn.click(ocr_all, inputs=[img_in, txt_in], outputs=[out_easy, out_paddle, out_tess])
|
| 60 |
|
| 61 |
+
demo.launch()
|