Research Repository

Assessing Capsule Networks with Biased Data

Ferrarini, Bruno and Ehsan, Shoaib and Bartoli, Adrien and Leonardis, Aleš and McDonald-Maier, Klaus D (2019) Assessing Capsule Networks with Biased Data. In: SCIA 2019 Scandinavian Conference on Image Analysis, 2019-06-11 - 2019-06-13, Norrköping, Sweden.

[img]
Preview
Text
1904.04555v1.pdf - Accepted Version

Download (1MB) | Preview

Abstract

Machine learning based methods achieves impressive results in object classification and detection. Utilizing representative data of the visual world during the training phase is crucial to achieve good performance with such data driven approaches. However, it not always possible to access bias-free datasets thus, robustness to biased data is a desirable property for a learning system. Capsule Networks have been introduced recently and their tolerance to biased data has received little attention. This paper aims to fill this gap and proposes two experimental scenarios to assess the tolerance to imbalanced training data and to determine the generalization performance of a model with unfamiliar affine transformations of the images. This paper assesses dynamic routing and EM routing based Capsule Networks and proposes a comparison with Convolutional Neural Networks in the two tested scenarios. The presented results provide new insights into the behaviour of capsule networks.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Published proceedings: Image Analysis 21st Scandinavian Conference, SCIA 2019, Norrköping, Sweden, June 11–13, 2019, Proceedings. Part of the Lecture Notes in Computer Science book series (LNCS, volume 11482). Also part of the Image Processing, Computer Vision, Pattern Recognition, and Graphics book sub series (LNIP, volume 11482)
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Depositing User: Elements
Date Deposited: 29 Apr 2020 15:12
Last Modified: 12 May 2020 01:00
URI: http://repository.essex.ac.uk/id/eprint/27393

Actions (login required)

View Item View Item