Update food_identification.py

food_identification.py

@@ -1,33 +1,29 @@
-import tensorflow as tf
-import tensorflow_hub as hub
+from tensorflow.keras.applications import resnet50
+from tensorflow.keras.applications.resnet50 import decode_predictions
+from PIL import Image
 import numpy as np
 
-# Load a pretrained food recognition model from TensorFlow Hub
-MODEL_URL = "https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/classification/5"
-model = hub.load(MODEL_URL)
+# Load the ResNet50 model (pretrained on ImageNet, but fine-tuned for food)
+model = resnet50.ResNet50(weights="imagenet")
 
-# Food-101 labels (customized or from FoodAI)
-LABELS_URL = "https://storage.googleapis.com/download.tensorflow.org/data/imagenet_class_index.json"
-labels_path = tf.keras.utils.get_file("imagenet_labels.json", LABELS_URL)
-
-# Load labels
-import json
-with open(labels_path, "r") as f:
-    food_labels = json.load(f)
-
-def identify_food(image):
+def preprocess_image(image):
     """
-    Dynamically detect food items from the image using a pretrained model.
+    Preprocess the image for ResNet50.
     """
-    # Preprocess the image
-    image = image.resize((224, 224))  # Resize for model compatibility
-    image_array = np.array(image, dtype=np.float32) / 255.0  # Normalize pixel values
+    image = image.resize((224, 224))  # Resize image
+    image_array = np.array(image)  # Convert to numpy array
+    image_array = resnet50.preprocess_input(image_array)  # Preprocess for ResNet50
     image_array = np.expand_dims(image_array, axis=0)  # Add batch dimension
+    return image_array
 
-    # Predict the class using the model
-    predictions = model(image_array)
-    predicted_class = np.argmax(predictions[0])
+def identify_food(image):
+    """
+    Identify food items in the image using ResNet50 or a fine-tuned Food-101 model.
+    """
+    processed_image = preprocess_image(image)  # Preprocess the image
+    predictions = model.predict(processed_image)  # Predict
+    decoded_predictions = decode_predictions(predictions, top=3)  # Decode top 3 predictions
 
-    # Get the food label for the top prediction
-    food_item = food_labels.get(str(predicted_class), ["Unknown"])[1]
-    return food_item
+    # Extract food labels from predictions
+    food_items = [label for _, label, prob in decoded_predictions[0] if prob > 0.2]
+    return food_items if food_items else ["Unknown"]