Domain Shift in Capsule Networks

Rajath S., Sumukh K., S. Natarajan

Abstract

Capsule Networks are an exciting deep learning architecture which overcomes some of the shortcomings of Convolutional Neural Networks (CNNs). Capsule networks aim to capture spatial relationships between parts of an object and exhibits viewpoint invariance. In practical computer vision, the training data distribution is different from the test distribution and the covariate shift affects the performance of the model. This problem is called Domain Shift. In this paper, we analyze how well capsule networks adapt to new domains by experimenting with multiple routing algorithms and comparing it with CNNs.

Download


Paper Citation


in Harvard Style

S. R., K. S. and Natarajan S. (2021). Domain Shift in Capsule Networks.In Proceedings of the 10th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-486-2, pages 275-278. DOI: 10.5220/0010252002750278


in Bibtex Style

@conference{icpram21,
author={Rajath S. and Sumukh K. and S. Natarajan},
title={Domain Shift in Capsule Networks},
booktitle={Proceedings of the 10th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2021},
pages={275-278},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010252002750278},
isbn={978-989-758-486-2},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 10th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Domain Shift in Capsule Networks
SN - 978-989-758-486-2
AU - S. R.
AU - K. S.
AU - Natarajan S.
PY - 2021
SP - 275
EP - 278
DO - 10.5220/0010252002750278