Update app.py

app.py

@@ -3,20 +3,45 @@ import numpy as np
 import cv2
 from tensorflow.keras.models import load_model
 
+IMG_HEIGHT = 96
+IMG_WIDTH = 96
+
 # Load the saved Keras model
-model = load_model("path_to_your_model.h5")  # Replace with the path to your ASL model
+model = load_model("model_01.keras")
 
 # Define the labels for ASL classes
-labels = ['A', 'B', 'C', 'D', 'E', 'F', ...]  # Replace with your actual label names
+labels = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I',
+         'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S',
+         'T', 'U', 'V', 'W', 'X', 'Y']  # Replace with your actual label names
 
 def preprocess_frame(frame):
-    """Preprocess the frame for the ASL model."""
-    # Resize to the input size expected by the model
-    img = cv2.resize(frame, (224, 224))  # Replace (224, 224) with your model's input size
-    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)  # Convert to RGB
-    img = img / 255.0  # Normalize pixel values
-    img = np.expand_dims(img, axis=0)  # Add batch dimension
-    return img
+    """Preprocess the video frame for the ASL model."""
+    # Convert the frame to a TensorFlow tensor
+    if isinstance(frame, np.ndarray):
+        frame = tf.convert_to_tensor(frame, dtype=tf.float32)
+        # Reshape to add channel dimension if grayscale
+        if frame.ndim == 2:  # If the input is grayscale
+            frame = tf.expand_dims(frame, axis=-1)
+            frame = tf.image.grayscale_to_rgb(frame)
+
+    # Ensure the frame has 3 channels (RGB)
+    if frame.shape[-1] == 1:  # Grayscale image
+        frame = tf.image.grayscale_to_rgb(frame)
+
+    # First scale down to dataset dimensions (if applicable)
+    frame = tf.image.resize(frame, [28, 28])  # Resize to smaller dimensions for consistency
+
+    # Resize to the target model input dimensions
+    frame = tf.image.resize(frame, [IMG_HEIGHT, IMG_WIDTH])
+
+    # Normalize pixel values to [0, 1]
+    frame = tf.cast(frame, tf.float32) / 255.0
+
+    # Add batch dimension for model input
+    frame = tf.expand_dims(frame, axis=0)
+
+    return frame
+
 
 def predict_asl(frame):
     """Predict the ASL sign from the webcam frame."""