Parallel Capsule Networks for Classification of White Blood Cells
This work addresses the challenge of improving CapsNet performance for medical image classification, specifically for leukemia diagnosis, representing an incremental advancement in domain-specific applications.
The authors tackled the problem of Capsule Networks (CapsNets) underperforming on complex datasets by proposing a parallel CapsNet architecture that branches to isolate capsules, improving classification of white blood cells in an unbalanced acute myeloid leukemia image dataset. Their results showed that parallel CapsNets outperformed ResNeXt-50 in accuracy, stability, and rotational invariance, whereas conventional CapsNets were unstable and similar to the baseline.
Capsule Networks (CapsNets) is a machine learning architecture proposed to overcome some of the shortcomings of convolutional neural networks (CNNs). However, CapsNets have mainly outperformed CNNs in datasets where images are small and/or the objects to identify have minimal background noise. In this work, we present a new architecture, parallel CapsNets, which exploits the concept of branching the network to isolate certain capsules, allowing each branch to identify different entities. We applied our concept to the two current types of CapsNet architectures, studying the performance for networks with different layers of capsules. We tested our design in a public, highly unbalanced dataset of acute myeloid leukaemia images (15 classes). Our experiments showed that conventional CapsNets show similar performance than our baseline CNN (ResNeXt-50) but depict instability problems. In contrast, parallel CapsNets can outperform ResNeXt-50, is more stable, and shows better rotational invariance than both, conventional CapsNets and ResNeXt-50.