University of Bahrain
Scientific Journals

Overview of CapsNet Performance Evaluation Methods for Image Classification using a Dual Input Capsule Network as a Case Study

Show simple item record Mensah, Patrick Kwabena Ayidzoe, Mighty Abra 2022-02-12T01:01:13Z 2022-02-12T01:01:13Z 2022-02-15
dc.identifier.issn 2210-142X
dc.description.abstract Performance evaluation is a critical part of deep learning (DL) that requires careful conduct to enhance confidence and reliability. Several metrics exist to evaluate DL models, however, choosing one for a given model is not trivial, since it is not a one-fit-all solution. Practically, accuracy is the most popularly used evaluation metric for capsule networks (CapsNets). This is problematic for sensitive applications (e.g. health), since accuracy is overly optimistic in the presence of class imbalance, and does not permit the exact reporting of a model’s risk of bias and potential usefulness. This paper, therefore, aims at demonstrating the usefulness of other metrics for performance evaluation as well as interpretability through the implementation of a custom capsule model. The metrics are effective in measuring the real performance of the models in terms of accuracy (93.03% for proposed model), number of parameters ( ≈ 4 million fewer for proposed model), ability to scale and fail-safe, and the effectiveness of the routing process when evaluated on the datasets. Evaluating a CapsNet model with all these metrics has the potential to enhance the practitioner’s confidence and also improve model understandability and reliability. en_US
dc.language.iso en en_US
dc.publisher University Of Bahrain en_US
dc.subject Capsule networks en_US
dc.subject Deep learning en_US
dc.subject Performance evaluation en_US
dc.subject COVID-19 en_US
dc.subject Explainable artificial intelligence en_US
dc.title Overview of CapsNet Performance Evaluation Methods for Image Classification using a Dual Input Capsule Network as a Case Study en_US
dc.volume 11 en_US
dc.issue 1 en_US
dc.pagestart 29 en_US
dc.pageend 43 en_US
dc.contributor.authorcountry Ghana en_US
dc.contributor.authorcountry China en_US
dc.contributor.authoraffiliation Department of Computer Science and Informatics, University of Energy and Natural Resources en_US
dc.contributor.authoraffiliation School of Information and Software Engineering,University of Electronic Science and Technology of China en_US
dc.source.title International Journal of Computing and Digital Systems en_US
dc.abbreviatedsourcetitle IJCDS en_US

Files in this item

This item appears in the following Issue(s)

Show simple item record

All Journals

Advanced Search


Administrator Account