Update app.py

app.py

@@ -4,6 +4,8 @@ import numpy as np
 from PIL import Image
 import cv2
 import tensorflow as tf
+import base64
+import io
 
 model = keras.models.load_model('my_model (2).h5')
 
@@ -32,7 +34,20 @@ def make_gradcam_heatmap(img_array, model):
     return heatmap.numpy()
 
 def predict(input_image):
-    img = Image.fromarray(input_image).convert('RGB')
+    # Handle both file upload and base64 string
+    if isinstance(input_image, str) and input_image.startswith('data:'):
+        # Base64 input
+        base64_data = input_image.split(',')[1]
+        image_bytes = base64.b64decode(base64_data)
+        img = Image.open(io.BytesIO(image_bytes)).convert('RGB')
+    elif isinstance(input_image, str):
+        # Plain base64 without prefix
+        image_bytes = base64.b64decode(input_image)
+        img = Image.open(io.BytesIO(image_bytes)).convert('RGB')
+    else:
+        # Normal image array from Gradio
+        img = Image.fromarray(input_image).convert('RGB')
+    
     img = img.resize((224, 224))
     img_array = np.array(img) / 255.0
     img_array = np.expand_dims(img_array, axis=0)
@@ -68,7 +83,8 @@ demo = gr.Interface(
         gr.Textbox(label="Classification Result")
     ],
     title="Bone Cancer Detection (Osteosarcoma)",
-    description="Upload an H&E stained histopathology image to classify as Non-Tumor, Non-Viable-Tumor, Viable, or viable: non-viable."
+    description="Upload an H&E stained histopathology image to classify.",
+    api_name="predict"
 )
 
 demo.launch()