Academic Journal
C3Net: A cross-modal collaborative calibration of features for object detection using frames and events.
| Τίτλος: | C3Net: A cross-modal collaborative calibration of features for object detection using frames and events. |
|---|---|
| Συγγραφείς: | Chen Y; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China., Zhong J; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China., Guo Y; School of Electronic Information, Wuhan University, Wuhan, China., Xie Z; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China., Xiao J; School of Electronic Information, Wuhan University, Wuhan, China. Electronic address: xiaojs@whu.edu.cn., Chen P; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China. Electronic address: phchen@gdut.edu.cn. |
| Πηγή: | Neural networks : the official journal of the International Neural Network Society [Neural Netw] 2026 Jul; Vol. 199, pp. 108651. Date of Electronic Publication: 2026 Feb 02. |
| Τύπος έκδοσης: | Journal Article |
| Γλώσσα: | English |
| Στοιχεία περιοδικού: | Publisher: Pergamon Press Country of Publication: United States NLM ID: 8805018 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1879-2782 (Electronic) Linking ISSN: 08936080 NLM ISO Abbreviation: Neural Netw Subsets: MEDLINE |
| Imprint Name(s): | Original Publication: New York : Pergamon Press, [c1988- |
| Ιατρικοί όροι (MeSH): | Image Processing, Computer-Assisted* , Detection Algorithms*, Calibration |
| Περίληψη: | Object detection by fusing RGB frames and event streams is challenging due to their inherent heterogeneity and significant statistical disparities, which often lead to suboptimal fusion in existing methods. To address this, we introduce C3Net, a novel framework built upon a paradigm shift from direct feature merging to Collaborative Calibration. First, we propose an Adaptive Balancing Time Surface (ABTS) to generate motion-robust event representations by mitigating spatial inconsistencies caused by varying object velocities. Second, the core Cross-Modal Feature Collaborative Calibration Module (CM-FCCM) performs mutual calibration of RGB and event features across channel and spatial dimensions, reducing modality discrepancies before fusion; the calibrated features are then fed back to the respective backbones for enriched feature learning. Finally, an Adaptive Channel Fusion Module (ACFM) dynamically integrates the modalities based on channel-wise confidence. Extensive experiments on PKU-DAVIS-SOD, DSEC-MOD, and PKU-DDD17-CAR datasets demonstrate that C3Net achieves state-of-the-art performance, showcasing its superior ability to leverage the complementary strengths of frames and events. (Copyright © 2026 Elsevier Ltd. All rights reserved.) |
| Competing Interests: | Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Yunhua Chen reports financial support was provided by Department of Science and Technology of Guangdong Province. Pinghua Chen reports financial support was provided by Department of Science and Technology of Guangdong Province. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. |
| Contributed Indexing: | Keywords: Event camera; Multimodal fusion; Object detection |
| Entry Date(s): | Date Created: 20260212 Date Completed: 20260827 Latest Revision: 20260827 |
| Update Code: | 20260827 |
| DOI: | 10.1016/j.neunet.2026.108651 |
| PMID: | 41679046 |
| Βάση Δεδομένων: | MEDLINE |
| FullText | Links: – Type: other Url: https://resolver.ebsco.com:443/public/rma-ftfapi/ejs/direct?AccessToken=4BE5BFED58F023AACF47&Show=Object Text: Availability: 0 CustomLinks: – Url: https://www.doi.org/10.1016/j.neunet.2026.108651? Name: ScienceDirect (all content) (s7799221) Category: fullText Text: View record from ScienceDirect MouseOverText: View record from ScienceDirect |
|---|---|
| Header | DbId: cmedm DbLabel: MEDLINE An: 41679046 AccessLevel: 3 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: C3Net: A cross-modal collaborative calibration of features for object detection using frames and events. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AU" term="%22Chen+Y%22">Chen Y</searchLink>; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China.<br /><searchLink fieldCode="AU" term="%22Zhong+J%22">Zhong J</searchLink>; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China.<br /><searchLink fieldCode="AU" term="%22Guo+Y%22">Guo Y</searchLink>; School of Electronic Information, Wuhan University, Wuhan, China.<br /><searchLink fieldCode="AU" term="%22Xie+Z%22">Xie Z</searchLink>; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China.<br /><searchLink fieldCode="AU" term="%22Xiao+J%22">Xiao J</searchLink>; School of Electronic Information, Wuhan University, Wuhan, China. Electronic address: xiaojs@whu.edu.cn.<br /><searchLink fieldCode="AU" term="%22Chen+P%22">Chen P</searchLink>; School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China. Electronic address: phchen@gdut.edu.cn. – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%228805018%22">Neural networks : the official journal of the International Neural Network Society</searchLink> [Neural Netw] 2026 Jul; Vol. 199, pp. 108651. <i>Date of Electronic Publication: </i>2026 Feb 02. – Name: TypePub Label: Publication Type Group: TypPub Data: Journal Article – Name: Language Label: Language Group: Lang Data: English – Name: TitleSource Label: Journal Info Group: Src Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22Pergamon+Press%22">Pergamon Press </searchLink><i>Country of Publication: </i>United States <i>NLM ID: </i>8805018 <i>Publication Model: </i>Print-Electronic <i>Cited Medium: </i>Internet <i>ISSN: </i>1879-2782 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%2208936080%22">08936080 </searchLink><i>NLM ISO Abbreviation: </i>Neural Netw <i>Subsets: </i>MEDLINE – Name: PublisherInfo Label: Imprint Name(s) Group: PubInfo Data: <i>Original Publication</i>: New York : Pergamon Press, [c1988- – Name: SubjectMESH Label: MeSH Terms Group: Su Data: <searchLink fieldCode="MM" term="%22Image+Processing%2C+Computer-Assisted%22">Image Processing, Computer-Assisted*</searchLink> <br /><searchLink fieldCode="MM" term="%22Detection+Algorithms%22">Detection Algorithms*</searchLink><br /><searchLink fieldCode="MH" term="%22Calibration%22">Calibration</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Object detection by fusing RGB frames and event streams is challenging due to their inherent heterogeneity and significant statistical disparities, which often lead to suboptimal fusion in existing methods. To address this, we introduce C3Net, a novel framework built upon a paradigm shift from direct feature merging to Collaborative Calibration. First, we propose an Adaptive Balancing Time Surface (ABTS) to generate motion-robust event representations by mitigating spatial inconsistencies caused by varying object velocities. Second, the core Cross-Modal Feature Collaborative Calibration Module (CM-FCCM) performs mutual calibration of RGB and event features across channel and spatial dimensions, reducing modality discrepancies before fusion; the calibrated features are then fed back to the respective backbones for enriched feature learning. Finally, an Adaptive Channel Fusion Module (ACFM) dynamically integrates the modalities based on channel-wise confidence. Extensive experiments on PKU-DAVIS-SOD, DSEC-MOD, and PKU-DDD17-CAR datasets demonstrate that C3Net achieves state-of-the-art performance, showcasing its superior ability to leverage the complementary strengths of frames and events.<br /> (Copyright © 2026 Elsevier Ltd. All rights reserved.) – Name: Abstract Label: Competing Interests Group: Ab Data: Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Yunhua Chen reports financial support was provided by Department of Science and Technology of Guangdong Province. Pinghua Chen reports financial support was provided by Department of Science and Technology of Guangdong Province. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. – Name: SubjectMinor Label: Contributed Indexing Group: Data: <i>Keywords: </i>Event camera; Multimodal fusion; Object detection – Name: DateEntry Label: Entry Date(s) Group: Date Data: <i>Date Created: </i>20260212 <i>Date Completed: </i>20260827 <i>Latest Revision: </i>20260827 – Name: DateUpdate Label: Update Code Group: Date Data: 20260827 – Name: DOI Label: DOI Group: ID Data: 10.1016/j.neunet.2026.108651 – Name: AN Label: PMID Group: ID Data: 41679046 |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=41679046 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1016/j.neunet.2026.108651 Languages: – Code: eng Text: English PhysicalDescription: Pagination: StartPage: 108651 Subjects: – SubjectFull: Calibration Type: general – SubjectFull: Image Processing, Computer-Assisted Type: general – SubjectFull: Detection Algorithms Type: general Titles: – TitleFull: C3Net: A cross-modal collaborative calibration of features for object detection using frames and events. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Chen Y – PersonEntity: Name: NameFull: Zhong J – PersonEntity: Name: NameFull: Guo Y – PersonEntity: Name: NameFull: Xie Z – PersonEntity: Name: NameFull: Xiao J – PersonEntity: Name: NameFull: Chen P IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 07 Text: 2026 Jul Type: published Y: 2026 Identifiers: – Type: issn-electronic Value: 1879-2782 Numbering: – Type: volume Value: 199 Titles: – TitleFull: Neural networks : the official journal of the International Neural Network Society Type: main |
| ResultId | 1 |