import cv2 import numpy as np import tensorflow as tf from huggingface_hub import from_pretrained_keras from tensorflow.keras.optimizers import Adam from .constants import LEARNING_RATE def get_model(): """ Download the model from the Hugging Face Hub and compile it. """ model = from_pretrained_keras("keras-io/video-vision-transformer") model.compile( optimizer=Adam(learning_rate=LEARNING_RATE), loss="sparse_categorical_crossentropy", # metrics=[ # keras.metrics.SparseCategoricalAccuracy(name="accuracy"), # keras.metrics.SparseTopKCategoricalAccuracy(5, name="top-5-accuracy"), # ], ) return model model = get_model() labels = ['liver', 'kidney-right', 'kidney-left', 'femur-right', 'femur-left', 'bladder', 'heart', 'lung-right', 'lung-left', 'spleen', 'pancreas'] def predict_label(path): frames = load_video(path) dataloader = prepare_dataloader(frames) prediction = model.predict(dataloader)[0] label = np.argmax(prediction, axis=0) label = labels[label] return label def load_video(path): """ Load video from path and return a list of frames. The video is converted to grayscale because it is the format expected by the model. """ cap = cv2.VideoCapture(path) frames = [] try: while True: ret, frame = cap.read() if not ret: break frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) frames.append(frame) finally: cap.release() return np.array(frames) def prepare_dataloader(video): video = tf.expand_dims(video, axis=0) dataset = tf.data.Dataset.from_tensor_slices((video, np.array([0]))) dataloader = ( dataset.map(preprocess, num_parallel_calls=tf.data.AUTOTUNE) .batch(1) .prefetch(tf.data.AUTOTUNE) ) return dataloader @tf.function def preprocess(frames: tf.Tensor, label: tf.Tensor): """Preprocess the frames tensors and parse the labels.""" # Preprocess images frames = tf.image.convert_image_dtype( frames[ ..., tf.newaxis ], # The new axis is to help for further processing with Conv3D layers tf.float32, ) # Parse label label = tf.cast(label, tf.float32) return frames, label