Spaces:
Sleeping
Sleeping
app.py
Browse files
app.py
CHANGED
|
@@ -80,8 +80,7 @@ def classify_ecg(image):
|
|
| 80 |
logits = outputs.logits # Get the logits
|
| 81 |
predicted_class = torch.argmax(logits, dim=1).item() # Get the index of the highest score
|
| 82 |
|
| 83 |
-
# Return
|
| 84 |
-
return "Clear" if predicted_class == 1 else "Unclear" # Adjust based on your model's labels
|
| 85 |
except Exception as e:
|
| 86 |
st.error(f"Error during model inference: {str(e)}")
|
| 87 |
return "Error in classification."
|
|
@@ -102,13 +101,14 @@ def validate_ecg(image):
|
|
| 102 |
if mean_pixel_value > 200: # Adjust threshold based on your requirements
|
| 103 |
return False, "Image is too bright."
|
| 104 |
|
| 105 |
-
# Additional checks could be implemented here
|
| 106 |
return True, ""
|
| 107 |
|
| 108 |
-
def
|
| 109 |
draw = ImageDraw.Draw(image)
|
| 110 |
-
|
| 111 |
-
|
|
|
|
|
|
|
| 112 |
return image
|
| 113 |
|
| 114 |
# Streamlit UI for ECG Upload
|
|
@@ -125,13 +125,13 @@ if uploaded_file:
|
|
| 125 |
else:
|
| 126 |
st.image(image, caption="Uploaded ECG Image.", use_column_width=True)
|
| 127 |
# Make predictions using the Hugging Face model
|
| 128 |
-
|
| 129 |
-
if
|
| 130 |
-
|
| 131 |
-
st.image(
|
| 132 |
-
st.write(f"Predicted heart status:
|
| 133 |
else:
|
| 134 |
-
st.write(f"Predicted heart status:
|
| 135 |
|
| 136 |
# Text input for additional questions
|
| 137 |
user_question = st.text_input("Ask any question about your ECG result", key="user_question")
|
|
|
|
| 80 |
logits = outputs.logits # Get the logits
|
| 81 |
predicted_class = torch.argmax(logits, dim=1).item() # Get the index of the highest score
|
| 82 |
|
| 83 |
+
return predicted_class # Return the class index for further processing
|
|
|
|
| 84 |
except Exception as e:
|
| 85 |
st.error(f"Error during model inference: {str(e)}")
|
| 86 |
return "Error in classification."
|
|
|
|
| 101 |
if mean_pixel_value > 200: # Adjust threshold based on your requirements
|
| 102 |
return False, "Image is too bright."
|
| 103 |
|
|
|
|
| 104 |
return True, ""
|
| 105 |
|
| 106 |
+
def highlight_issue(image):
|
| 107 |
draw = ImageDraw.Draw(image)
|
| 108 |
+
width, height = image.size
|
| 109 |
+
|
| 110 |
+
# Draw a red line across the middle of the image to indicate an issue
|
| 111 |
+
draw.line([(0, height // 2), (width, height // 2)], fill="red", width=5)
|
| 112 |
return image
|
| 113 |
|
| 114 |
# Streamlit UI for ECG Upload
|
|
|
|
| 125 |
else:
|
| 126 |
st.image(image, caption="Uploaded ECG Image.", use_column_width=True)
|
| 127 |
# Make predictions using the Hugging Face model
|
| 128 |
+
predicted_class = classify_ecg(image)
|
| 129 |
+
if predicted_class == 0: # Assuming 0 means "Unclear"
|
| 130 |
+
highlighted_image = highlight_issue(image.copy())
|
| 131 |
+
st.image(highlighted_image, caption="Marked ECG Image with Issues.", use_column_width=True)
|
| 132 |
+
st.write(f"Predicted heart status: Unclear. Potential issue detected.")
|
| 133 |
else:
|
| 134 |
+
st.write(f"Predicted heart status: Clear.")
|
| 135 |
|
| 136 |
# Text input for additional questions
|
| 137 |
user_question = st.text_input("Ask any question about your ECG result", key="user_question")
|