Spaces:
Sleeping
Sleeping
| import tensorflow as tf | |
| from tensorflow.keras.applications.vgg16 import preprocess_input | |
| from tensorflow.keras.applications.inception_v3 import preprocess_input as inception_preprocess_input | |
| import matplotlib.pyplot as plt | |
| def retina_preprocessing(image_path): | |
| """ | |
| Preprocesses a retina image file. | |
| Args: | |
| image_path: a string, the path to the image file | |
| Returns: | |
| a tensor of shape (1, 224, 224, 3), the preprocessed image | |
| """ | |
| image = plt.imread(image_path) | |
| # If the image is 2D (grayscale), add a channel dimension | |
| if len(image.shape) == 2: | |
| image = tf.expand_dims(image, axis=-1) | |
| image = tf.image.grayscale_to_rgb(image) | |
| image = tf.image.resize(image, [224, 224]) | |
| image = preprocess_input(image) | |
| return tf.expand_dims(image, axis=0) | |
| def pneumonia_preprocessing(image_path): | |
| """ | |
| Preprocesses a chest X-ray image file. | |
| Args: | |
| image_path: a string, the path to the image file | |
| Returns: | |
| a tensor of shape (1, 224, 224, 3), the preprocessed image | |
| """ | |
| image = plt.imread(image_path) | |
| # If the image is 2D (grayscale), add a channel dimension | |
| if len(image.shape) == 2: | |
| image = tf.expand_dims(image, axis=-1) | |
| image = tf.image.grayscale_to_rgb(image) | |
| image = tf.image.resize(image, [224, 224]) | |
| image = tf.cast(image, tf.float32) / 255.0 | |
| return tf.expand_dims(image, axis=0) | |
| def kidney_brain_preprocessing(image_path): | |
| """ | |
| Preprocesses a kidney image file. | |
| Args: | |
| image_path: str, the path to the kidney image file | |
| Returns: | |
| tensor: a tensor of shape (1, 224, 224, 1), the preprocessed kidney image | |
| """ | |
| image = plt.imread(image_path) | |
| # If the image is 2D (grayscale), add a channel dimension | |
| if len(image.shape) == 2: | |
| image = tf.expand_dims(image, axis=-1) | |
| image = tf.image.resize(image, [224, 224]) | |
| # if not grayscale convert it | |
| if image.shape[-1] != 1: | |
| image = tf.image.rgb_to_grayscale(image) | |
| return tf.expand_dims(image, axis=0) | |
| def general_preprocessing(image_path): | |
| image = plt.imread(image_path) | |
| # If the image is 2D (grayscale), add a channel dimension | |
| if len(image.shape) == 2: | |
| image = tf.expand_dims(image, axis=-1) | |
| image = tf.image.grayscale_to_rgb(image) | |
| image = tf.image.resize(image, [224, 224]) | |
| image = inception_preprocess_input(image) | |
| return tf.expand_dims(image, axis=0) |