A4 Article in conference proceedings
Efficient CNN with uncorrelated Bag of Features pooling (2022)

Laakom, F., Raitoharju, J., Iosifidis, A., & Gabbouj, M. (2022). Efficient CNN with uncorrelated Bag of Features pooling. In SSCI 2022 : Proceedings of the 2022 IEEE Symposium Series on Computational Intelligence (pp. 1082-1087). IEEE. https://doi.org/10.1109/SSCI51031.2022.10022157

JYU authors or editors

Publication details

All authors or editors: Laakom, Firas; Raitoharju, Jenni; Iosifidis, Alexandros; Gabbouj, Moncef

Parent publication: SSCI 2022 : Proceedings of the 2022 IEEE Symposium Series on Computational Intelligence

Place and date of conference: Singapore, 4.-7.12.2022

ISBN: 978-1-6654-8769-6

eISBN: 978-1-6654-8768-9

Publication year: 2022

Publication date: 30/01/2023

Pages range: 1082-1087

Publisher: IEEE

Publication country: United States

Publication language: English

DOI: https://doi.org/10.1109/SSCI51031.2022.10022157

Publication open access: Not open

Publication channel open access: 

Web address of parallel published publication (pre-print): https://arxiv.org/abs/2209.10865

Abstract

Despite the superior performance of CNN, deploying them on low computational power devices is still limited as they are typically computationally expensive. One key cause of the high complexity is the connection between the convolution layers and the fully connected layers, which typically requires a high number of parameters. To alleviate this issue, Bag of Features (BoF) pooling has been recently proposed. BoF learns a dictionary, that is used to compile a histogram representation of the input. In this paper, we propose an approach that builds on top of BoF pooling to boost its efficiency by ensuring that the items of the learned dictionary are non-redundant. We propose an additional loss term, based on the pair-wise correlation of the items of the dictionary, which complements the standard loss to explicitly regularize the model to learn a more diverse and rich dictionary. The proposed strategy yields an efficient variant of BoF and further boosts its performance, without any additional parameters.

Keywords: computer vision; automated pattern recognition; machine learning; deep learning; neural networks (information technology)

Free keywords: deep learning; CNN; diversity; bag of features pooling

Contributing organizations

Related projects

Ministry reporting: Yes

VIRTA submission year: 2022

JUFO rating: 1