Initial Commit

app.py

@@ -5,6 +5,9 @@ import tensorflow as tf
 import mlxtend
 import joblib
 import cv2
+import matplotlib.pyplot as plt
+import seaborn as sns
+import tempfile
 from sklearn.preprocessing import LabelEncoder
 from tensorflow.keras.models import save_model, load_model
 from PIL import Image
@@ -58,6 +61,35 @@ ensemble_models = {
 class_labels = ['COVID-19', 'Normal', 'Viral Pneumonia']
 class_index_mapping = {i: label for i, label in enumerate(class_labels)}
 
+# Create a bar chart to represent symptom weightage
+def show_symptom_weightage(symptoms, weights):
+    plt.figure(figsize=(10, 6))
+    plt.bar(symptoms, weights)
+    plt.xlabel("Symptom")
+    plt.ylabel("Weight")
+    plt.title("Weightage/Importance of Symptoms")
+    plt.xticks(rotation=45, ha='right')
+    plt.tight_layout()
+    # Save the plot as an image and return the file path
+    file_path = tempfile.NamedTemporaryFile(suffix=".png").name
+    plt.savefig(file_path)
+    plt.close()
+    return file_path
+
+# Create a heat map to represent image-based model predictions
+def show_heat_map(image, prediction_probabilities):
+    plt.figure(figsize=(8, 6))
+    sns.heatmap(prediction_probabilities, annot=True, cmap="YlGnBu", xticklabels=class_labels, yticklabels=False)
+    plt.title("Image-based Model Prediction Heatmap")
+    plt.xlabel("Prediction")
+    plt.ylabel("Model")
+    plt.tight_layout()
+    # Save the plot as an image and return the file path
+    file_path = tempfile.NamedTemporaryFile(suffix=".png").name
+    plt.savefig(file_path)
+    plt.close()
+    return file_path
+    
 # Apply the same rescaling as in the model
 def rescale_images(img):
     return img / 127.5 - 1
@@ -82,7 +114,7 @@ def predict_covid(*args):
     image = args[-1]
 
     if not any(symptoms) and image is None:
-        return "COVID-19 Negative", "100%", "COVID-19 Negative", "100%", "No inputs defined"    
+        return "COVID-19 Negative", "100%", "COVID-19 Negative", "100%", "No inputs defined", None, None    
         
     # Prepare the input data for the text-based model
     label_encoder = LabelEncoder()
@@ -135,6 +167,9 @@ def predict_covid(*args):
 
     text_prob_positive = np.average(probabilities, weights=weights) * 100
 
+    # Visualize the symptom weightage/importance    
+    symptom_weightage_img = show_symptom_weightage(column_names[:-1], weights)
+    
     # Prepare the input data for the image-based model
     image_prediction = None
     confidence = None
@@ -148,7 +183,9 @@ def predict_covid(*args):
                 "NA",
                 "Invalid X-Ray Scan",
                 "NA",
-                "Your X-Ray scan does not meet the required standards. Please ensure that your scans are not blurry, pixelated, or disfigured"
+                "Your X-Ray scan does not meet the required standards. Please ensure that your scans are not blurry, pixelated, or disfigured",
+                None,
+                None
             )
         
         image = preprocess_image(image)
@@ -168,6 +205,9 @@ def predict_covid(*args):
         # Calculate the confidence level of the prediction
         confidence = np.max(avg_ensemble_prob) * 100
 
+        # Visualize the heatmap
+        heatmap_img = show_heat_map(image, [prediction for prediction, _ in ensemble_predictions])
+        
         # Provide reasoning for the prediction
         if ensemble_prediction == 0 and text_prediction == "COVID-19 Positive":
             reason = "Your X-Ray scan and symptoms provide conclusive indications of COVID-19" 
@@ -195,7 +235,7 @@ def predict_covid(*args):
     else:
         reason = "Your symptoms provide conclusive indications of your diagnosis"
         
-    return text_prediction,  f"{text_prob_positive:.2f}%", image_prediction, f"{confidence:.2f}%", reason
+    return text_prediction, f"{text_prob_positive:.2f}%", image_prediction, f"{confidence:.2f}%", reason, heatmap_img, symptom_weightage_img
 
 # Create the input and output components
 symptom_components = [gr_components.Checkbox(label=label) for label in column_names[:-1]]
@@ -206,6 +246,8 @@ output_components = [
     gr_components.Label(label="Prediction based on X-Ray Scan"),
     gr_components.Label(label="X-Ray Scan Confidence (%)"),
     gr_components.Textbox(label="Final Prediction"),
+    gr_components.Image(label="X-Ray Prediction Heatmap"),
+    gr_components.Image(label="Weightage assigned to Symptoms")
 ]
 
 # Create the interface and launch