update app

app.py

@@ -9,27 +9,79 @@ processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
 # Define the BMI classes
 bmi_classes = ["underweight", "normal weight", "overweight", "obesity"]
 
+# Define the secondary models for BMI predictions and midpoints
+bmi_ranges = {
+    "underweight": {
+        "BMI < 16.0": (0, 16.0, 8.0),
+        "16.0 ≤ BMI ≤ 16.99": (16.0, 16.99, 16.5),
+        "17.0 ≤ BMI ≤ 18.49": (17.0, 18.49, 17.75)
+    },
+    "normal weight": {
+        "18.5 ≤ BMI ≤ 20.4": (18.5, 20.4, 19.45),
+        "20.5 ≤ BMI ≤ 22.4": (20.5, 22.4, 21.45),
+        "22.5 ≤ BMI ≤ 24.9": (22.5, 24.9, 23.7)
+    },
+    "overweight": {
+        "25.0 ≤ BMI ≤ 26.9": (25.0, 26.9, 25.95),
+        "27.0 ≤ BMI ≤ 28.9": (27.0, 28.9, 27.95),
+        "29.0 ≤ BMI ≤ 29.9": (29.0, 29.9, 29.45)
+    },
+    "obesity": {
+        "30.0 ≤ BMI ≤ 34.9": (30.0, 34.9, 32.5),
+        "35.0 ≤ BMI ≤ 39.9": (35.0, 39.9, 37.45),
+        "BMI ≥ 40.0": (40.0, 100, 40.0)  # Assuming 100 as the upper limit for BMI
+    }
+}
+
 def predict_bmi(image):
     # Prepare the inputs
     inputs = processor(text=bmi_classes, images=image, return_tensors="pt", padding=True)
     # Get model outputs
     outputs = model(**inputs)
-    logits_per_image = outputs.logits_per_image # Image-text similarity scores
-    probs = logits_per_image.softmax(dim=1) # Convert to probabilities
+    logits_per_image = outputs.logits_per_image  # Image-text similarity scores
+    probs = logits_per_image.softmax(dim=1)  # Convert to probabilities
 
     # Find the class with the highest probability
     max_prob_index = probs.argmax().item()
     predicted_bmi_class = bmi_classes[max_prob_index]
 
-    return predicted_bmi_class
+    # Use the midpoint BMI for the predicted class
+    bmi_prediction = get_midpoint_bmi(predicted_bmi_class)
+    
+    # Assume height is input by the user or extracted somehow
+    height_in_inches = 75  # Example height; replace with actual input or extraction
+    predicted_weight = calculate_weight(bmi_prediction, height_in_inches)
+
+    # Create the JSON output
+    result = {
+        "weightCategory": f"{predicted_bmi_class} - {bmi_prediction}",
+        "bmiPrediction": f"{bmi_prediction:.2f}",
+        "height": str(height_in_inches),
+        "predictedWeight": f"{predicted_weight:.2f} lbs"
+    }
+
+    return result
+
+def get_midpoint_bmi(weight_category):
+    """Return the midpoint BMI for the given weight category."""
+    category_ranges = bmi_ranges.get(weight_category.lower())
+    for range_label, (low, high, mid) in category_ranges.items():
+        return mid  # Return the first midpoint found in the given range
+
+def calculate_weight(bmi, height_in_inches):
+    """Calculate the weight from BMI and height (in inches)."""
+    height_in_meters = height_in_inches * 0.0254  # Convert height to meters
+    weight_kg = bmi * (height_in_meters ** 2)  # BMI formula to calculate weight
+    weight_lbs = weight_kg * 2.20462  # Convert kg to lbs
+    return weight_lbs
 
-# Create Gradio interface
+# Create Gradio interface with updated components
 interface = gr.Interface(
     fn=predict_bmi,
-    inputs=gr.inputs.Image(type="pil"),
-    outputs="text",
+    inputs=gr.Image(type="pil"),
+    outputs="json",
     title="BMI Prediction",
-    description="Upload an image to predict BMI category (underweight, normal weight, overweight, obesity)."
+    description="Upload an image to predict BMI category (underweight, normal weight, overweight, obesity) and receive a detailed prediction."
 )
 
 # Launch the interface