RT Journal Article SR Electronic T1 Combining Attention-based Multiple Instance Learning and Gaussian Processes for CT Hemorrhage Detection JF bioRxiv FD Cold Spring Harbor Laboratory SP 2021.07.01.450539 DO 10.1101/2021.07.01.450539 A1 Yunan Wu A1 Arne Schmidt A1 Enrique Hernández-Sánchez A1 Rafael Molina A1 Aggelos K. Katsaggelos YR 2021 UL http://biorxiv.org/content/early/2021/07/04/2021.07.01.450539.abstract AB Intracranial hemorrhage (ICH) is a life-threatening emergency with high rates of mortality and morbidity. Rapid and accurate detection of ICH is crucial for patients to get a timely treatment. In order to achieve the automatic diagnosis of ICH, most deep learning models rely on huge amounts of slice labels for training. Unfortunately, the manual annotation of CT slices by radiologists is time-consuming and costly. To diagnose ICH, in this work, we propose to use an attention-based multiple instance learning (Att-MIL) approach implemented through the combination of an attention-based convolutional neural network (Att-CNN) and a variational Gaussian process for multiple instance learning (VGP-MIL). Only labels at scan-level are necessary for training. Our method trains the model using scan labels and assigns each slice with an attention weight, which can be used to provide slice-level predictions, and uses the VGPMIL model based on low-dimensional features extracted by the Att-CNN to obtain improved predictions both at slice and scan levels. To analyze the performance of the proposed approach, our model has been trained on 1150 scans from an RSNA dataset and evaluated on 490 scans from an external CQ500 dataset. Our method outperforms other methods using the same scan-level training and is able to achieve comparable or even better results than other methods relying on slicelevel annotations.Competing Interest StatementThe authors have declared no competing interest.