Abstract: Techniques for diagnosing a patient having an AAA. The techniques include using a computer hardware processor to perform: accessing computed tomography angiography (CTA) images of a portion of the patient, the portion of the patient including the AAA of the patient; providing the CTA images as input to a trained machine learning model, the trained machine learning model being configured to classify a property of the AAA based on the CTA images; and determining, based on the classified property of the AAA, a diagnosis of the patient, the diagnosis including information identifying at least one condition (e.g., an endoleak, an AAA rupture) of the patient.

Abstract: A method and system for domain adaptation based on multi-layer fusion in a convolutional neural network architecture for feature extraction and a two-step training and fine-tuning scheme. The architecture concatenates features extracted at different depths of the network to form a fully connected layer before the classification step. First, the network is trained with a large set of images from a source domain as a feature extractor. Second, for each new domain (including the source domain), the classification step is fine-tuned with images collected from the corresponding site. The features from different depths are concatenated with and fine-tuned with weights adjusted for a specific task. The architecture is used for classifying high occupancy vehicle images.

Abstract: A method and system for domain adaptation based on multi-layer fusion in a convolutional neural network architecture for feature extraction and a two-step training and fine-tuning scheme. The architecture concatenates features extracted at different depths of the network to form a fully connected layer before the classification step. First, the network is trained with a large set of images from a source domain as a feature extractor. Second, for each new domain (including the source domain), the classification step is fine-tuned with images collected from the corresponding site. The features from different depths are concatenated with and fine-tuned with weights adjusted for a specific task. The architecture is used for classifying high occupancy vehicle images.