Update app.py

app.py

@@ -0,0 +1,65 @@
+import os
+import cv2
+import numpy as np
+import tensorflow as tf
+import pickle
+import tkinter as tk
+from tkinter import filedialog
+
+# Set global parameters (update these as per your model's training settings)
+IMAGE_WIDTH = 128
+IMAGE_HEIGHT = 64
+MAX_TEXT_LENGTH = 32
+
+# Function to preprocess the image
+def preprocess_image(image_path, image_width, image_height): 
+    img = cv2.imread(image_path, cv2.IMREAD_GRAYSCALE)
+    if img is None:
+        print(f"Error reading image: {image_path}")
+        return None
+    img = cv2.resize(img, (image_width, image_height))
+    img = img / 255.0
+    img = img.T  # Transpose if your model requires it
+    img = np.expand_dims(img, axis=-1)
+    img = np.expand_dims(img, axis=0)
+    return img
+
+# Function to predict the word from the image using your model and tokenizer
+def predict_word(model, image_path, tokenizer, max_text_length):
+    img = preprocess_image(image_path, IMAGE_WIDTH, IMAGE_HEIGHT)
+    if img is None:
+        return ""
+    
+    # Get the model prediction
+    prediction = model.predict(img)
+    predicted_indices = np.argmax(prediction, axis=-1)[0]
+    
+    # Create a reverse mapping from index to character
+    reverse_index = {index: char for char, index in tokenizer.word_index.items()}
+    
+    # Convert prediction indices into a string (ignoring padding zeros)
+    predicted_word = ''.join([reverse_index.get(index, '') for index in predicted_indices if index != 0])
+    return predicted_word
+
+# Load your trained Keras model
+model = tf.keras.models.load_model("handwritten_text_recognition_model.keras")
+
+# Load the tokenizer (make sure 'tokenizer.pickle' exists in your project directory)
+with open("tokenizer.pickle", "rb") as handle:
+    tokenizer = pickle.load(handle)
+
+# Create a simple Tkinter file dialog to select an image
+root = tk.Tk()
+root.withdraw()  # Hide the main window
+
+# Open file dialog for image selection
+image_path = filedialog.askopenfilename(
+    title="Select an Image", 
+    filetypes=[("Image Files", "*.png;*.jpg;*.jpeg")]
+)
+
+if image_path:
+    predicted_word = predict_word(model, image_path, tokenizer, MAX_TEXT_LENGTH)
+    print("Predicted word:", predicted_word)
+else:
+    print("No image selected.")