<abstract xmlns="http://www.w3.org/1999/xhtml"><p>The topical question studied in this paper is how many receptive fields (filters) a convolutional layer of a convolutional neural network should have. The goal is to find a rule for choosing the most appropriate numbers of filters. The benchmark datasets are principally diverse CIFAR-10 and EEACL26 to use a common network architecture with three convolutional layers whose numbers of filters are changeable. Heterogeneity and sensitiveness of CIFAR-10 with infiniteness and scalability of EEACL26 are believed to be relevant enough for generalization and spreading of the appropriateness of filter numbers. The appropriateness rule is drawn from top accuracies obtained on 10 × 20 × 21 parallelepipeds for three image sizes. They show, knowing that the number of filters of the first convolutional layer should be set greater for the more complex dataset, the rest of appropriate numbers of filters are set at integers, which are multiples of that number. The multipliers make a sequence similar to a progression, e.g., it may be 1, 3, 9, 15 or 1, 2, 8, 16, etc. With only those multipliers, such a rule-of-progression does not give the number of filters for the first convolutional layer.</p></abstract>

The topical question studied in this paper is how many receptive fields (filters) a convolutional layer of a convolutional neural network should have. The goal is to find a rule for choosing the most appropriate numbers of filters. The benchmark datasets are principally diverse CIFAR-10 and EEACL26 to use a common network architecture with three convolutional layers whose numbers of filters are changeable. Heterogeneity and sensitiveness of CIFAR-10 with infiniteness and scalability of EEACL26 are believed to be relevant enough for generalization and spreading of the appropriateness of filter numbers. The appropriateness rule is drawn from top accuracies obtained on 10 × 20 × 21 parallelepipeds for three image sizes. They show, knowing that the number of filters of the first convolutional layer should be set greater for the more complex dataset, the rest of appropriate numbers of filters are set at integers, which are multiples of that number. The multipliers make a sequence similar to a progression, e.g., it may be 1, 3, 9, 15 or 1, 2, 8, 16, etc. With only those multipliers, such a rule-of-progression does not give the number of filters for the first convolutional layer.

Appropriateness of Numbers of Receptive Fields in Convolutional Neural Networks Based on Classifying CIFAR-10 and EEACL26 Datasets

Electrical, Control and Communication Engineering

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

The topical question studied in this paper is how many receptive fields (filters) a convolutional layer of a convolutional neural network should have. The goal...

Appropriateness of Numbers of Receptive Fields in Convolutional Neural Networks Based on Classifying CIFAR-10 and EEACL26 Datasets

Publié en ligne: 12 mars 2019

Pages: 157 - 163

DOI: https://doi.org/10.2478/ecce-2018-0019

Mots clés
Convolutional neural networks, Convolutional layers, Filters, Performance, Receptive fields

© 2018 Vadim Romanuke, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Appropriateness of Numbers of Receptive Fields in Convolutional Neural Networks Based on Classifying CIFAR-10 and EEACL26 Datasets

Publié en ligne: 12 mars 2019

Pages: 157 - 163

DOI: https://doi.org/10.2478/ecce-2018-0019

Mots clésConvolutional neural networks, Convolutional layers, Filters, Performance, Receptive fields

© 2018 Vadim Romanuke, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Mots clés
Convolutional neural networks, Convolutional layers, Filters, Performance, Receptive fields