Update app.py

app.py

@@ -5,58 +5,59 @@ import numpy as np
 import tensorflow as tf
 import gradio as gr
 
-# 1. Load your trained model
-model = tf.keras.models.load_model("best_model.h5")
+# Load your trained model without trying to restore the old optimizer/loss,
+# then re-compile it so we can call predict()
+model = tf.keras.models.load_model("best_model.h5", compile=False)
+model.compile(optimizer="adam", loss="mse")
 
-# 2. Denoising pipeline function
 def process_and_denoise(image):
     """
-    Input: HxW grayscale or color image (numpy array).
-    Outputs: (orig, noisy, denoised) all as 64×64 uint8 arrays.
+    Takes any input image (color or grayscale), converts it to 64×64 grayscale,
+    adds Gaussian noise, and returns:
+      1) the resized original
+      2) the noisy version
+      3) the model’s denoised reconstruction
     """
-    # --- Preprocess Original ---
-    # If color, convert to gray:
+    # If color, convert to gray; otherwise accept 1-channel
     if image.ndim == 3 and image.shape[2] == 3:
         gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
     else:
-        # could be (H,W,1) or (H,W)
-        gray = image[...,0] if image.ndim == 3 else image
+        # Could be (H,W,1) or (H,W)
+        gray = image[..., 0] if image.ndim == 3 else image
 
-    # Resize to 64×64
-    orig = cv2.resize(gray, (64,64))
+    # Resize and normalize
+    orig = cv2.resize(gray, (64, 64))
     orig_norm = orig.astype(np.float32) / 255.0
 
-    # --- Add Gaussian Noise ---
-    sigma = 0.1  # same noise level you trained with
+    # Add Gaussian noise
+    sigma = 0.1
     noisy = orig_norm + sigma * np.random.randn(*orig_norm.shape)
     noisy = np.clip(noisy, 0.0, 1.0)
 
-    # --- Denoise with the autoencoder ---
-    inp = noisy[np.newaxis, ..., np.newaxis]   # shape (1,64,64,1)
-    pred = model.predict(inp)[0, ..., 0]       # shape (64,64)
+    # Denoise via the autoencoder
+    inp = noisy[np.newaxis, ..., np.newaxis]    # shape (1,64,64,1)
+    pred = model.predict(inp)[0, ..., 0]        # shape (64,64)
 
-    # --- Convert all to uint8 for display ---
-    orig_disp   = (orig_norm * 255).astype(np.uint8)
-    noisy_disp  = (noisy    * 255).astype(np.uint8)
-    recon_disp  = (pred     * 255).astype(np.uint8)
+    # Convert back to uint8 for display
+    orig_disp = (orig_norm  * 255).astype(np.uint8)
+    noisy_disp = (noisy      * 255).astype(np.uint8)
+    recon_disp = (pred       * 255).astype(np.uint8)
 
     return orig_disp, noisy_disp, recon_disp
 
-# 3. Build Gradio interface with 3 outputs
 demo = gr.Interface(
     fn=process_and_denoise,
     inputs=gr.Image(type="numpy", label="Input Image"),
     outputs=[
         gr.Image(type="numpy", label="Original (64×64)"),
         gr.Image(type="numpy", label="Noisy (σ=0.1)"),
-        gr.Image(type="numpy", label="Reconstructed")
+        gr.Image(type="numpy", label="Denoised Output")
     ],
     title="Denoising Autoencoder Demo",
     description=(
         "Upload any image (grayscale or color). "
-        "This app converts it to 64×64 grayscale, adds Gaussian noise, "
-        "then denoises it with the trained autoencoder. "
-        "See all three side by side!"
+        "This will convert it to 64×64 grayscale, add Gaussian noise, "
+        "and then denoise it with the trained autoencoder."
     )
 )