real-time-parking-occupancy-detection icon indicating copy to clipboard operation
real-time-parking-occupancy-detection copied to clipboard

extract_features_from_images.py where is this file plz?

Open tiyarocks opened this issue 4 years ago • 1 comments

tiyarocks avatar Dec 09 '20 10:12 tiyarocks

From this notebook in "[23]"

ground_truth = []
prediction = []
for img in valid_images:
    label = img.split("/")[-2]
    if label == "Occupied":
        ground_truth.append(1)
    else:
        ground_truth.append(0)
    image = load_image(img)
    image = cv2.resize(image, (WIDTH, HEIGHT))
    image_x = np.expand_dims(image, axis=0)
    image_x = preprocess_input(image_x)
    pred = tf_model.predict(image_x)
    pred = np.squeeze(pred)
    if pred > 0.98:
        prediction.append(1)
    else:
        prediction.append(0)

You can create valid dataset in this notebook with the code below.

valid_y = []
valid_x = []
for img in valid_images:
    label = img.split("/")[-2]
    if label == "Occupied":
        valid_y.append(1)
    else:
        valid_y.append(0)
    image = load_image(img)
    image = cv2.resize(image, (WIDTH, HEIGHT))
    image_x = np.expand_dims(image, axis=0)
    image_x = preprocess_input(image_x)
    pred = tf_model.predict(image_x)
    pred = np.squeeze(pred)
    valid_x.append(pred)
valid_x = np.array(valid_x)
valid_x = valid_x.reshape(-1, 1)
valid_y = np.array(valid_y)

This is the same way with the train dataset

train255 avatar Dec 28 '20 02:12 train255