Random sampling for deep learning segmentation of acute ischemic stroke on non-contrast CT

A method for generating segmentation masks to assist in identification of acute ischemic stroke, the method comprising:

• (a) performing a non-contrast computed tomography scan to produce a computed tomography image;
• (b) inputting the computed tomography image to an input layer of a deep learning neural network;
• (c) outputting a segmentation mask of acute ischemic stroke from an output layer of the deep learning neural network, wherein the segmentation mask of acute ischemic stroke is generated in response to the computed tomography image input to the deep learning neural network;
• wherein the deep learning neural network is trained with a training set comprising ground truth non-contrast computed tomography images and corresponding segmentation masks of acute ischemic stroke,
• wherein the training set comprises multiple segmentation masks of acute ischemic stroke for each of the non-contrast computed tomography images;
• and
• wherein one of the multiple segmentation masks is randomly and uniformly sampled from the multiple segmentation masks for training.