RIC-CNN: Rotation-Invariant Coordinate Convolutional Neural Network.
Saved in:
| Title: | RIC-CNN: Rotation-Invariant Coordinate Convolutional Neural Network. |
|---|---|
| Authors: | Mo, Hanlin1 (AUTHOR) hanlin.mo@oulu.fi, Zhao, Guoying1 (AUTHOR) guoying.zhao@oulu.fi |
| Source: | Pattern Recognition. Feb2024, Vol. 146, pN.PAG-N.PAG. 1p. |
| Subjects: | Convolutional neural networks, Image recognition (Computer vision), Data augmentation, Remote sensing |
| Abstract: | Due to the lack of rotation invariance in traditional convolution operations, even acting a slight rotation on the input can severely degrade the performance of Convolutional Neural Networks (CNNs). To address this, we propose a Rotation-Invariant Coordinate Convolution (RIC-C), which achieves natural invariance to arbitrary rotations around the input center without additional trainable parameters or data augmentation. We first evaluate the rotational invariance of RIC-C using the MNIST dataset and compare its performance with most previous rotation-invariant CNN models. RIC-C achieves state-of-the-art classification on the MNIST-rot test set without data augmentation and with lower computational costs. Then, the interchangeability of RIC-C with traditional convolution operations is demonstrated by seamlessly integrating it into common CNN models like VGG, ResNet, and DenseNet. We conduct remote sensing image classification on the NWPU VHR-10, MTARSI and AID datasets and patch matching experiments on the UBC benchmark dataset, showing that RIC-C significantly enhances the performance of CNN models across different applications, especially when training data is limited. Our codes can be downloaded from https://github.com/HanlinMo/Rotation-Invariant-Coordinate-Convolutional-Neural-Network.git. • We propose RIC-C: a novel convolutional operation naturally invariant to all input center rotations, no extra parameters or data augmentation. • Without data augmentation, RIC-CNN shows superior performance on MNIST compared to previous rotation invariant CNNs. • RIC-C is successfully deployed to popular CNN models, and enhances their performance in various tasks, particularly with limited training data. [ABSTRACT FROM AUTHOR] |
| Copyright of Pattern Recognition is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
| FullText | Text: Availability: 0 |
|---|---|
| Header | DbId: egs DbLabel: Engineering Source An: 173416052 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: RIC-CNN: Rotation-Invariant Coordinate Convolutional Neural Network. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Mo%2C+Hanlin%22">Mo, Hanlin</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> hanlin.mo@oulu.fi</i><br /><searchLink fieldCode="AR" term="%22Zhao%2C+Guoying%22">Zhao, Guoying</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> guoying.zhao@oulu.fi</i> – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%22Pattern+Recognition%22">Pattern Recognition</searchLink>. Feb2024, Vol. 146, pN.PAG-N.PAG. 1p. – Name: Subject Label: Subjects Group: Su Data: <searchLink fieldCode="DE" term="%22Convolutional+neural+networks%22">Convolutional neural networks</searchLink><br /><searchLink fieldCode="DE" term="%22Image+recognition+%28Computer+vision%29%22">Image recognition (Computer vision)</searchLink><br /><searchLink fieldCode="DE" term="%22Data+augmentation%22">Data augmentation</searchLink><br /><searchLink fieldCode="DE" term="%22Remote+sensing%22">Remote sensing</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Due to the lack of rotation invariance in traditional convolution operations, even acting a slight rotation on the input can severely degrade the performance of Convolutional Neural Networks (CNNs). To address this, we propose a Rotation-Invariant Coordinate Convolution (RIC-C), which achieves natural invariance to arbitrary rotations around the input center without additional trainable parameters or data augmentation. We first evaluate the rotational invariance of RIC-C using the MNIST dataset and compare its performance with most previous rotation-invariant CNN models. RIC-C achieves state-of-the-art classification on the MNIST-rot test set without data augmentation and with lower computational costs. Then, the interchangeability of RIC-C with traditional convolution operations is demonstrated by seamlessly integrating it into common CNN models like VGG, ResNet, and DenseNet. We conduct remote sensing image classification on the NWPU VHR-10, MTARSI and AID datasets and patch matching experiments on the UBC benchmark dataset, showing that RIC-C significantly enhances the performance of CNN models across different applications, especially when training data is limited. Our codes can be downloaded from https://github.com/HanlinMo/Rotation-Invariant-Coordinate-Convolutional-Neural-Network.git. • We propose RIC-C: a novel convolutional operation naturally invariant to all input center rotations, no extra parameters or data augmentation. • Without data augmentation, RIC-CNN shows superior performance on MNIST compared to previous rotation invariant CNNs. • RIC-C is successfully deployed to popular CNN models, and enhances their performance in various tasks, particularly with limited training data. [ABSTRACT FROM AUTHOR] – Name: AbstractSuppliedCopyright Label: Group: Ab Data: <i>Copyright of Pattern Recognition is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.) |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=173416052 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1016/j.patcog.2023.109994 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 1 StartPage: N.PAG Subjects: – SubjectFull: Convolutional neural networks Type: general – SubjectFull: Image recognition (Computer vision) Type: general – SubjectFull: Data augmentation Type: general – SubjectFull: Remote sensing Type: general Titles: – TitleFull: RIC-CNN: Rotation-Invariant Coordinate Convolutional Neural Network. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Mo, Hanlin – PersonEntity: Name: NameFull: Zhao, Guoying IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 02 Text: Feb2024 Type: published Y: 2024 Identifiers: – Type: issn-print Value: 00313203 Numbering: – Type: volume Value: 146 Titles: – TitleFull: Pattern Recognition Type: main |
| ResultId | 1 |
