please

import numpy as np
import cv2
from tensorflow.keras.models import load_model

model = load_model('best_model_v4.h5')  # Load your trained model
target_size = (200, 200)  # Example target size, should match your model's input size

num_to_label = {
    0: 'A',
    1: 'B',
    2: 'C',
    3: 'D',
    4: 'E',
    5: 'F',
    6: 'G',
    7: 'H',
    8: 'I',
    9: 'J',
    10: 'K',
    11: 'L',
    12: 'M',
    13: 'N',
    14: 'O',
    15: 'P',
    16: 'Q',
    17: 'R',
    18: 'S',
    19: 'T',
    20: 'U',
    21: 'V',
    22: 'W',
    23: 'X',
    24: 'Y',
    25: 'Z',
    26: 'del',
    27: 'nothing',
    28: 'space'
}

def preprocess_frame(frame, target_size):
    frame = crop_to_square(frame)  # Crop the frame to a square
    frame = cv2.resize(frame, target_size)
    frame = frame / 255.0  # Normalize to [0, 1] range
    frame = frame.astype('float32')
    frame = np.expand_dims(frame, axis=0)  # Add batch dimension
    return frame

def crop_to_square(frame):
    h, w, _ = frame.shape
    if h > w:
        diff = (h - w) // 2
        frame = frame[diff:diff + w, :]
    else:
        diff = (w - h) // 2
        frame = frame[:, diff:diff + h]
    return frame

cap = cv2.VideoCapture(0)  # Start capturing video from the first webcam

while True:
    ret, frame = cap.read()
    if not ret:
        break

    # Crop the frame to a square before any processing
    square_frame = crop_to_square(frame)

    # Preprocess the frame
    processed_frame = preprocess_frame(frame, target_size)
    prediction = model.predict(processed_frame)
    
    # Decode the prediction (depends on your model's output format)
    predicted_label = np.argmax(prediction)

    # Map the predicted label to the corresponding letter
    predicted_label = num_to_label[predicted_label]
    
    # Display the label on the square frame
    cv2.putText(square_frame, str(predicted_label), (30, 50), cv2.FONT_HERSHEY_SIMPLEX, 2, (255, 0, 0), 2, cv2.LINE_AA)
    cv2.imshow('Live ASL Recognition', square_frame)
    
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()