Predicting the presence of a disease in volumetric images is an essential task in medical imaging. The use of state-of-the-art techniques like deep convolutional neural networks (CNN) for such tasks is challenging due to limited supervised training data and high memory usage. This paper presents a weakly supervised solution that can be used in learning deep CNN features for volumetric image classification. In the proposed method, we use extreme value theory to model the feature distribution of the images without a pathology and use it to identify positive instances in an image that contains pathology. The experimental results show that the proposed method can learn classifiers that have similar performance to a fully supervised method and have significantly better performance in comparison with methods that use fixed number of instances from a positive image.

Additional Metadata
Keywords Intra-retinal fluid, Learning (artificial intelligence), Macular Edema, Medical image processing, Multiple instance learning, OCT images, Optical coherence tomography, Pigment Epithelial Detachment, Retinal fluid classification, ReTOUCH challenge, Sub-retinal fluid, Weak supervision
Persistent URL dx.doi.org/10.1007/978-3-030-20893-6_37, hdl.handle.net/1765/117318
Series Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Note e-book; no subscription
Citation
Tennakoon, R.B, Gostar, A.K. (Amirali K.), Hoseinnezhad, R. (Reza), de Bruijne, M, & Bab-Hadiashar, A. (2019). Deep Multi-instance Volumetric Image Classification with Extreme Value Distributions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). doi:10.1007/978-3-030-20893-6_37