Cluster-CAM: Cluster-weighted visual interpretation of CNNs’ decision in image classification
Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN’s decision, has drawn increasing attention. Gradient-based C...
Saved in:
| Published in: | Neural networks Vol. 178; p. 106473 |
|---|---|
| Main Authors: | , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
United States
Elsevier Ltd
01.10.2024
|
| Subjects: | |
| ISSN: | 0893-6080, 1879-2782, 1879-2782 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN’s decision, has drawn increasing attention. Gradient-based CAMs are efficient, while the performance is heavily affected by gradient vanishing and exploding. In contrast, gradient-free CAMs can avoid computing gradients to produce more understandable results. However, they are quite time-consuming because hundreds of forward interference per image are required. In this paper, we proposed Cluster-CAM, an effective and efficient gradient-free CNN interpretation algorithm. Cluster-CAM can significantly reduce the times of forward propagation by splitting the feature maps into clusters. Furthermore, we propose an artful strategy to forge a cognition-base map and cognition-scissors from clustered feature maps. The final salience heatmap will be produced by merging the above cognition maps. Qualitative results conspicuously show that Cluster-CAM can produce heatmaps where the highlighted regions match the human’s cognition more precisely than existing CAMs. The quantitative evaluation further demonstrates the superiority of Cluster-CAM in both effectiveness and efficiency. |
|---|---|
| AbstractList | Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN's decision, has drawn increasing attention. Gradient-based CAMs are efficient, while the performance is heavily affected by gradient vanishing and exploding. In contrast, gradient-free CAMs can avoid computing gradients to produce more understandable results. However, they are quite time-consuming because hundreds of forward interference per image are required. In this paper, we proposed Cluster-CAM, an effective and efficient gradient-free CNN interpretation algorithm. Cluster-CAM can significantly reduce the times of forward propagation by splitting the feature maps into clusters. Furthermore, we propose an artful strategy to forge a cognition-base map and cognition-scissors from clustered feature maps. The final salience heatmap will be produced by merging the above cognition maps. Qualitative results conspicuously show that Cluster-CAM can produce heatmaps where the highlighted regions match the human's cognition more precisely than existing CAMs. The quantitative evaluation further demonstrates the superiority of Cluster-CAM in both effectiveness and efficiency.Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN's decision, has drawn increasing attention. Gradient-based CAMs are efficient, while the performance is heavily affected by gradient vanishing and exploding. In contrast, gradient-free CAMs can avoid computing gradients to produce more understandable results. However, they are quite time-consuming because hundreds of forward interference per image are required. In this paper, we proposed Cluster-CAM, an effective and efficient gradient-free CNN interpretation algorithm. Cluster-CAM can significantly reduce the times of forward propagation by splitting the feature maps into clusters. Furthermore, we propose an artful strategy to forge a cognition-base map and cognition-scissors from clustered feature maps. The final salience heatmap will be produced by merging the above cognition maps. Qualitative results conspicuously show that Cluster-CAM can produce heatmaps where the highlighted regions match the human's cognition more precisely than existing CAMs. The quantitative evaluation further demonstrates the superiority of Cluster-CAM in both effectiveness and efficiency. Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN's decision, has drawn increasing attention. Gradient-based CAMs are efficient, while the performance is heavily affected by gradient vanishing and exploding. In contrast, gradient-free CAMs can avoid computing gradients to produce more understandable results. However, they are quite time-consuming because hundreds of forward interference per image are required. In this paper, we proposed Cluster-CAM, an effective and efficient gradient-free CNN interpretation algorithm. Cluster-CAM can significantly reduce the times of forward propagation by splitting the feature maps into clusters. Furthermore, we propose an artful strategy to forge a cognition-base map and cognition-scissors from clustered feature maps. The final salience heatmap will be produced by merging the above cognition maps. Qualitative results conspicuously show that Cluster-CAM can produce heatmaps where the highlighted regions match the human's cognition more precisely than existing CAMs. The quantitative evaluation further demonstrates the superiority of Cluster-CAM in both effectiveness and efficiency. |
| ArticleNumber | 106473 |
| Author | Ji, Hongbing Stanković, Ljubiša Feng, Zhenpeng Zhu, Mingzhe Daković, Miloš Cui, Xiyang |
| Author_xml | – sequence: 1 givenname: Zhenpeng orcidid: 0000-0002-0383-4794 surname: Feng fullname: Feng, Zhenpeng organization: School of Electronic Engineering, Xidian University, Xi’an, China – sequence: 2 givenname: Hongbing surname: Ji fullname: Ji, Hongbing email: hbji@xidian.edu.cn organization: School of Electronic Engineering, Xidian University, Xi’an, China – sequence: 3 givenname: Miloš orcidid: 0000-0002-3317-3632 surname: Daković fullname: Daković, Miloš organization: Faculty of Electrical Engineering, University of Montenegro, Podgorica, Montenegro – sequence: 4 givenname: Xiyang surname: Cui fullname: Cui, Xiyang organization: School of Electronic Engineering, Xidian University, Xi’an, China – sequence: 5 givenname: Mingzhe orcidid: 0000-0002-7962-3344 surname: Zhu fullname: Zhu, Mingzhe organization: School of Electronic Engineering, Xidian University, Xi’an, China – sequence: 6 givenname: Ljubiša orcidid: 0000-0002-9736-9036 surname: Stanković fullname: Stanković, Ljubiša organization: Faculty of Electrical Engineering, University of Montenegro, Podgorica, Montenegro |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/38941740$$D View this record in MEDLINE/PubMed |
| BookMark | eNqFkMtKxDAUhoMoOo6-gUiXbjrm1ibjQpDiDbxsdB3a5FQzdNoxSRV3voav55OYsY4LFwqBkJ_vP5x822i97VpAaI_gCcEkP5xNWuhbCBOKKY9RzgVbQyMixTSlQtJ1NMJyytIcS7yFtr2fYYxzydkm2mJyyongeIRU0fQ-gEuLk-ujZPV4AfvwGMAkz9b3ZZPYNqYLB6EMtmuTrk6Kmxv_8faeGNDWLzMbz7x8gEQ3pfe2tvqL3UEbddl42P2-x-j-7PSuuEivbs8vi5OrVFMhQkoErqgRjHKoasEpw4aRquIYsgwybUT8nRYAxvA6k5oZTvmUESmFxlBxzcboYJi7cN1TDz6oufUamqZsoeu9YliwPKOS5xHd_0b7ag5GLVxc3L2qlZQI8AHQrvPeQf2DEKyW7tVMDe7V0r0a3Mfa0a-atoOw4Erb_Fc-HsoQJT1bcMprC60GYx3ooExn_x7wCXzHou4 |
| CitedBy_id | crossref_primary_10_1007_s00521_025_10978_0 crossref_primary_10_1016_j_eswa_2025_127238 crossref_primary_10_3390_land13101585 crossref_primary_10_3390_pr13051365 crossref_primary_10_1016_j_asoc_2025_113265 |
| Cites_doi | 10.1016/j.neunet.2021.09.018 10.1016/j.neucom.2022.09.129 10.1007/978-3-031-19775-8_27 10.1109/ICCV.2017.444 10.1016/j.neunet.2022.10.035 10.1016/j.neucom.2022.11.084 10.1016/j.neucom.2022.05.065 10.1016/j.neucom.2021.03.056 10.1016/j.neunet.2022.08.009 10.3390/rs13091772 10.3390/rs13204139 10.1109/TNNLS.2019.2944672 10.1109/CVPR.2016.308 10.1016/j.neucom.2019.05.043 10.1109/MSP.2019.2929832 10.1109/ICCV48922.2021.00137 10.1109/CVPRW50498.2020.00020 10.1109/ICCV.2017.74 10.1016/j.sigpro.2021.108301 10.1109/CVPR52688.2022.01167 10.1016/j.neucom.2022.09.029 10.1109/ICCV.2019.00980 10.1038/s41467-019-08987-4 10.1016/j.sigpro.2022.108685 10.1109/TNNLS.2021.3066850 10.1109/CVPR.2016.319 10.1109/CVPR46437.2021.01625 10.1109/MSP.2017.2696572 |
| ContentType | Journal Article |
| Copyright | 2024 Elsevier Ltd Copyright © 2024 Elsevier Ltd. All rights reserved. |
| Copyright_xml | – notice: 2024 Elsevier Ltd – notice: Copyright © 2024 Elsevier Ltd. All rights reserved. |
| DBID | AAYXX CITATION NPM 7X8 |
| DOI | 10.1016/j.neunet.2024.106473 |
| DatabaseName | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitle | CrossRef PubMed MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic PubMed |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| EISSN | 1879-2782 |
| ExternalDocumentID | 38941740 10_1016_j_neunet_2024_106473 S0893608024003976 |
| Genre | Journal Article |
| GroupedDBID | --- --K --M -~X .DC .~1 0R~ 123 186 1B1 1RT 1~. 1~5 29N 4.4 457 4G. 53G 5RE 5VS 6TJ 7-5 71M 8P~ 9JM 9JN AABNK AACTN AAEDT AAEDW AAIKJ AAKOC AALRI AAOAW AAQFI AAQXK AAXLA AAXUO AAYFN ABAOU ABBOA ABCQJ ABEFU ABFNM ABFRF ABHFT ABIVO ABJNI ABLJU ABMAC ABXDB ACDAQ ACGFO ACGFS ACIUM ACNNM ACRLP ACZNC ADBBV ADEZE ADGUI ADJOM ADMUD ADRHT AEBSH AECPX AEFWE AEKER AENEX AFKWA AFTJW AFXIZ AGHFR AGUBO AGWIK AGYEJ AHHHB AHJVU AHZHX AIALX AIEXJ AIKHN AITUG AJOXV AKRWK ALMA_UNASSIGNED_HOLDINGS AMFUW AMRAJ AOUOD ARUGR ASPBG AVWKF AXJTR AZFZN BJAXD BKOJK BLXMC CS3 DU5 EBS EFJIC EJD EO8 EO9 EP2 EP3 F0J F5P FDB FEDTE FGOYB FIRID FNPLU FYGXN G-2 G-Q G8K GBLVA GBOLZ HLZ HMQ HVGLF HZ~ IHE J1W JJJVA K-O KOM KZ1 LG9 LMP M2V M41 MHUIS MO0 MOBAO MVM N9A O-L O9- OAUVE OZT P-8 P-9 P2P PC. Q38 R2- RIG ROL RPZ SBC SCC SDF SDG SDP SES SEW SNS SPC SPCBC SSN SST SSV SSW SSZ T5K TAE UAP UNMZH VOH WUQ XPP ZMT ~G- 9DU AATTM AAXKI AAYWO AAYXX ABDPE ABWVN ACLOT ACRPL ACVFH ADCNI ADNMO AEIPS AEUPX AFJKZ AFPUW AGQPQ AIGII AIIUN AKBMS AKYEP ANKPU APXCP CITATION EFKBS EFLBG ~HD BNPGV NPM SSH 7X8 |
| ID | FETCH-LOGICAL-c277t-170b2d7324ebf74230d31bb40e55e5cd7647c7eedd4f58c3d424931887c0eb4c3 |
| ISSN | 0893-6080 1879-2782 |
| IngestDate | Sun Sep 28 15:23:20 EDT 2025 Thu Apr 03 07:07:43 EDT 2025 Tue Nov 18 21:57:24 EST 2025 Sat Nov 29 05:33:08 EST 2025 Sat Aug 10 15:30:49 EDT 2024 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Keywords | Explainable artificial intelligence Clustering algorithm Class activation mapping Image classification |
| Language | English |
| License | Copyright © 2024 Elsevier Ltd. All rights reserved. |
| LinkModel | OpenURL |
| MergedId | FETCHMERGED-LOGICAL-c277t-170b2d7324ebf74230d31bb40e55e5cd7647c7eedd4f58c3d424931887c0eb4c3 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0002-9736-9036 0000-0002-0383-4794 0000-0002-7962-3344 0000-0002-3317-3632 |
| PMID | 38941740 |
| PQID | 3073652846 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_3073652846 pubmed_primary_38941740 crossref_primary_10_1016_j_neunet_2024_106473 crossref_citationtrail_10_1016_j_neunet_2024_106473 elsevier_sciencedirect_doi_10_1016_j_neunet_2024_106473 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-10-01 |
| PublicationDateYYYYMMDD | 2024-10-01 |
| PublicationDate_xml | – month: 10 year: 2024 text: 2024-10-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | Neural networks |
| PublicationTitleAlternate | Neural Netw |
| PublicationYear | 2024 |
| Publisher | Elsevier Ltd |
| Publisher_xml | – name: Elsevier Ltd |
| References | Chattopadhay, Sarkar, Howlader, Balasubramanian (b2) 2018 (pp. 618–626). Ramaswamy, H. G., et al. (2020). Ablation-cam: Visual explanations for deep convolutional network via gradient-free localization. In Zhu, Feng, Stanković, Ding, Fan, Zhou (b44) 2022 (pp. 1316–1324). Saleem, Yuan, Kurugollu, Anjum, Liu (b23) 2022; 513 Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2016). Learning deep features for discriminative localization. In (pp. 983–991). Sun, Song, Cai, Du, Guizani (b31) 2022; 500 Feng, Zhu, Stanković, Ji (b6) 2021; 13 He, Zhang, Ren, Sun (b8) 2016 Jung, H., & Oh, Y. (2021). Towards Better Explanations of Class Activation Mapping. In Chen, Jin, Jin, Zhu, Chen (b3) 2022; 33 Scalzo, Stanković, Daković, Constantinides, Mandic (b24) 2023; 158 Fu, R., Hu, Q., Dong, X., Guo, Y., Gao, Y., & Li, B. (2020). Axiom-based Grad-CAM: Towards accurate visualization and explanation of CNNs. In Townsend, Chaton, Monteiro (b34) 2020; 31 Wang, H., Wang, Z., Du, M., Yang, F., Zhang, Z., Ding, S., et al. (2020). Score-CAM: Score-weighted visual explanations for convolutional neural networks. In Zhu, Y., Zhao, C., Wang, J., Zhao, X., Wu, Y., & Lu, H. (2017). Couplenet: Coupling global structure with local parts for object detection. In Tan, Gao, Khan, Guan (b33) 2022; 155 Zeiler, Fergus (b38) 2014 Ma, Zhang, Pena-Pena, Arce (b16) 2021; 189 Feng, Ji, Stanković, Fan, Zhu (b5) 2021; 13 (pp. 4126–4134). . Redmon, Divvala, Girshick, Farhadi (b21) 2016 Zhang, Rao, Yang (b39) 2021 Krizhevsky, Sutskever, Hinton (b10) 2012 Stankovic, Mandic, Dakovic, Kisil, Sejdic, Constantinides (b30) 2019; 36 (pp. 24–25). (pp. 459–474). Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017). Grad-CAM: Visual explanations from deep networks via gradient-based localization. In Lapuschkin, Wäldchen, Binder, Montavon, Samek, Müller (b11) 2019; 10 (pp. 2921–2929). (pp. 9705–9714). Zheng, Q., Wang, Z., Zhou, J., & Lu, J. (2022). Shap-CAM: Visual explanations for convolutional neural networks based on Shapley Value. In Zhou, Kainz (b42) 2018 Simonyan, K., & Zisserman, A. (2015). Very Deep Convolutional Networks for Large-Scale Image Recognition. In Cao, J., Pang, Y., Han, J., & Li, X. (2019). Hierarchical shot detector. In Petsiuk, Das, Saenko (b19) 2018 Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the Inception Architecture for Computer Vision. In Liu, Zhang, Zhou, Wang (b15) 2023; 521 (pp. 1–14). Ren, Li, Liu, Zhang (b22) 2021 Stankovic, Dakovic, Sejdic (b29) 2017; 34 Spinelli, Scardapane, Uncini (b27) 2022 Liu, Meng, Li, Mao, Chen (b14) 2022; 510 Tu, Zhou, Gan, Jiang, Hussain, Luo (b35) 2021; 449 Liang, Hu, Zhang, Lin, Xing (b12) 2018; 31 Vlahek, Mongus (b36) 2021 (pp. 11976–11986). Srinivas, A., Lin, T.-Y., Parmar, N., Shlens, J., Abbeel, P., & Vaswani, A. (2021). Bottleneck transformers for visual recognition. In Omeiza, Speakman, Cintas, Weldermariam (b18) 2019 Macpherson, Churchland, Sejnowski, DiCarlo, Kamitani, Takahashi (b17) 2021; 144 (pp. 16519–16529). Zhao, Xie, Wang, Liu, Shi, Du (b40) 2019; 359 Liu, Z., Mao, H., Wu, C. Y., Feichtenhofer, C., Darrell, T., & Xie, S. (2022). A convnet for the 2020s. In Deng, Dong, Socher, Li, Li, Fei-Fei (b4) 2009 (pp. 2818–2826). Stankovic (10.1016/j.neunet.2024.106473_b30) 2019; 36 10.1016/j.neunet.2024.106473_b32 Zhang (10.1016/j.neunet.2024.106473_b39) 2021 Deng (10.1016/j.neunet.2024.106473_b4) 2009 Saleem (10.1016/j.neunet.2024.106473_b23) 2022; 513 Tu (10.1016/j.neunet.2024.106473_b35) 2021; 449 10.1016/j.neunet.2024.106473_b13 Macpherson (10.1016/j.neunet.2024.106473_b17) 2021; 144 10.1016/j.neunet.2024.106473_b37 Omeiza (10.1016/j.neunet.2024.106473_b18) 2019 Vlahek (10.1016/j.neunet.2024.106473_b36) 2021 Zhu (10.1016/j.neunet.2024.106473_b44) 2022 He (10.1016/j.neunet.2024.106473_b8) 2016 10.1016/j.neunet.2024.106473_b1 Chattopadhay (10.1016/j.neunet.2024.106473_b2) 2018 Lapuschkin (10.1016/j.neunet.2024.106473_b11) 2019; 10 Stankovic (10.1016/j.neunet.2024.106473_b29) 2017; 34 10.1016/j.neunet.2024.106473_b9 Zhao (10.1016/j.neunet.2024.106473_b40) 2019; 359 Ren (10.1016/j.neunet.2024.106473_b22) 2021 10.1016/j.neunet.2024.106473_b28 Feng (10.1016/j.neunet.2024.106473_b5) 2021; 13 10.1016/j.neunet.2024.106473_b7 Liang (10.1016/j.neunet.2024.106473_b12) 2018; 31 Zeiler (10.1016/j.neunet.2024.106473_b38) 2014 Chen (10.1016/j.neunet.2024.106473_b3) 2022; 33 10.1016/j.neunet.2024.106473_b41 10.1016/j.neunet.2024.106473_b20 10.1016/j.neunet.2024.106473_b43 10.1016/j.neunet.2024.106473_b45 Feng (10.1016/j.neunet.2024.106473_b6) 2021; 13 Ma (10.1016/j.neunet.2024.106473_b16) 2021; 189 Redmon (10.1016/j.neunet.2024.106473_b21) 2016 10.1016/j.neunet.2024.106473_b25 10.1016/j.neunet.2024.106473_b26 Krizhevsky (10.1016/j.neunet.2024.106473_b10) 2012 Zhou (10.1016/j.neunet.2024.106473_b42) 2018 Tan (10.1016/j.neunet.2024.106473_b33) 2022; 155 Liu (10.1016/j.neunet.2024.106473_b14) 2022; 510 Liu (10.1016/j.neunet.2024.106473_b15) 2023; 521 Sun (10.1016/j.neunet.2024.106473_b31) 2022; 500 Townsend (10.1016/j.neunet.2024.106473_b34) 2020; 31 Scalzo (10.1016/j.neunet.2024.106473_b24) 2023; 158 Spinelli (10.1016/j.neunet.2024.106473_b27) 2022 Petsiuk (10.1016/j.neunet.2024.106473_b19) 2018 |
| References_xml | – volume: 10 start-page: 1 year: 2019 end-page: 8 ident: b11 article-title: Unmasking clever hans predictors and assessing what machines really learn publication-title: Nature Communications – volume: 521 start-page: 27 year: 2023 end-page: 40 ident: b15 article-title: BFMNet: Bilateral feature fusion network with multi-scale context aggregation for real-time semantic segmentation publication-title: Neurocomputing – reference: (pp. 983–991). – reference: Srinivas, A., Lin, T.-Y., Parmar, N., Shlens, J., Abbeel, P., & Vaswani, A. (2021). Bottleneck transformers for visual recognition. In – reference: (pp. 2921–2929). – volume: 144 start-page: 603 year: 2021 end-page: 613 ident: b17 article-title: Natural and Artificial Intelligence: A brief introduction to the interplay between AI and neuroscience research publication-title: Neural Networks – reference: Cao, J., Pang, Y., Han, J., & Li, X. (2019). Hierarchical shot detector. In – volume: 510 start-page: 193 year: 2022 end-page: 202 ident: b14 article-title: SiSL-Net: Saliency-guided self-supervised learning network for image classification publication-title: Neurocomputing – volume: 449 start-page: 443 year: 2021 end-page: 454 ident: b35 article-title: A novel domain activation mapping-guided network (DA-GNT) for visual tracking publication-title: Neurocomputing – year: 2018 ident: b42 article-title: Efficient image evidence analysis of cnn classification results – volume: 31 year: 2018 ident: b12 article-title: Symbolic graph reasoning meets convolutions publication-title: Advances in Neural Information Processing Systems – reference: (pp. 2818–2826). – start-page: 8971 year: 2021 end-page: 8981 ident: b22 article-title: Interpreting and disentangling feature components of various complexity from DNNs publication-title: Proceedings of international conference on machine learning – reference: (pp. 4126–4134). – volume: 13 start-page: 1772 year: 2021 ident: b6 article-title: Self-matching CAM: A novel accurate visual explanation of CNNs for SAR image interpretation publication-title: Remote Sensing – start-page: 839 year: 2018 end-page: 847 ident: b2 article-title: Grad-CAM++: Generalized gradient-based visual explanations for deep convolutional networks publication-title: Proceedings of 2018 IEEE winter conference on applications of computer vision – reference: Fu, R., Hu, Q., Dong, X., Guo, Y., Gao, Y., & Li, B. (2020). Axiom-based Grad-CAM: Towards accurate visualization and explanation of CNNs. In – reference: Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the Inception Architecture for Computer Vision. In – year: 2018 ident: b19 article-title: RISE: Randomized input sampling for explanation of black-box models – year: 2021 ident: b39 article-title: Group-CAM: Group score-weighted visual explanations for deep convolutional networks – reference: Simonyan, K., & Zisserman, A. (2015). Very Deep Convolutional Networks for Large-Scale Image Recognition. In – reference: (pp. 1316–1324). – year: 2022 ident: b44 article-title: A probe-feature for specific emitter identification using axiom-based grad-CAM publication-title: Signal Processing – reference: Wang, H., Wang, Z., Du, M., Yang, F., Zhang, Z., Ding, S., et al. (2020). Score-CAM: Score-weighted visual explanations for convolutional neural networks. In – start-page: 818 year: 2014 end-page: 833 ident: b38 article-title: Visualizing and understanding convolutional networks publication-title: European conference on computer vision – reference: (pp. 11976–11986). – reference: Zhu, Y., Zhao, C., Wang, J., Zhao, X., Wu, Y., & Lu, H. (2017). Couplenet: Coupling global structure with local parts for object detection. In – volume: 158 start-page: 83 year: 2023 end-page: 88 ident: b24 article-title: A class of doubly stochastic shift operators for random graph signals and their boundedness publication-title: Neural Networks – reference: Liu, Z., Mao, H., Wu, C. Y., Feichtenhofer, C., Darrell, T., & Xie, S. (2022). A convnet for the 2020s. In – volume: 500 start-page: 989 year: 2022 end-page: 1002 ident: b31 article-title: CAMA: Class activation mapping disruptive attack for deep neural networks publication-title: Neurocomputing – volume: 31 start-page: 3456 year: 2020 end-page: 3470 ident: b34 article-title: Extracting relational explanations from deep neural networks: A survey from a neural-symbolic perspective publication-title: IEEE Transactions on Neural Networks and Learning Systems – reference: (pp. 16519–16529). – volume: 189 year: 2021 ident: b16 article-title: Fast spectral clustering method based on graph similarity matrix completion publication-title: Signal Processing – volume: 513 start-page: 165 year: 2022 end-page: 180 ident: b23 article-title: Explaining deep neural networks: A survey on the global interpretation methods publication-title: Neurocomputing – reference: Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017). Grad-CAM: Visual explanations from deep networks via gradient-based localization. In – year: 2019 ident: b18 article-title: Smooth grad-cam++: An enhanced inference level visualization technique for deep convolutional neural network models – start-page: 1 year: 2022 end-page: 9 ident: b27 article-title: A meta-learning approach for training explainable graph neural networks publication-title: IEEE Transactions on Neural Networks and Learning Systems – start-page: 248 year: 2009 end-page: 255 ident: b4 article-title: ImageNet: A large-scale hierarchical image database publication-title: Proceedings of 2009 IEEE conference on computer vision and pattern recognition – volume: 359 start-page: 185 year: 2019 end-page: 198 ident: b40 article-title: Visualizing and understanding of learned compressive sensing with residual network publication-title: Neurocomputing – volume: 34 start-page: 176 year: 2017 end-page: 182 ident: b29 article-title: Vertex-frequency analysis: A way to localize graph spectral components [lecture notes] publication-title: IEEE Signal Processing Magazine – reference: Jung, H., & Oh, Y. (2021). Towards Better Explanations of Class Activation Mapping. In – reference: (pp. 24–25). – volume: 33 start-page: 4991 year: 2022 end-page: 5003 ident: b3 article-title: Semisupervised semantic segmentation by improving prediction confidence publication-title: IEEE Transactions on Neural Networks and Learning Systems – reference: Ramaswamy, H. G., et al. (2020). Ablation-cam: Visual explanations for deep convolutional network via gradient-free localization. In – reference: (pp. 9705–9714). – reference: . – volume: 155 start-page: 58 year: 2022 end-page: 73 ident: b33 article-title: Interpretable artificial intelligence through locality guided neural networks publication-title: Neural Networks – start-page: 770 year: 2016 end-page: 778 ident: b8 article-title: Deep residual learning for image recognition publication-title: Proceedings of 2016 IEEE conference on computer vision and pattern recognition – reference: (pp. 618–626). – reference: Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2016). Learning deep features for discriminative localization. In – reference: (pp. 459–474). – reference: (pp. 1–14). – year: 2012 ident: b10 article-title: ImageNet classification with deep convolutional neural networks publication-title: Advances in neural information processing systems, vol. 25 – start-page: 779 year: 2016 end-page: 788 ident: b21 article-title: You only look once: Unified, real-time object detection publication-title: Proceedings of 2016 IEEE conference on computer vision and pattern recognition – reference: Zheng, Q., Wang, Z., Zhou, J., & Lu, J. (2022). Shap-CAM: Visual explanations for convolutional neural networks based on Shapley Value. In – volume: 36 start-page: 133 year: 2019 end-page: 145 ident: b30 article-title: Understanding the basis of graph signal processing via an intuitive example-driven approach publication-title: IEEE Signal Processing Magazine – volume: 13 start-page: 4139 year: 2021 ident: b5 article-title: SC-SM CAM: An efficient visual interpretation of CNN for SAR images target recognition publication-title: Remote Sensing – start-page: 1 year: 2021 end-page: 13 ident: b36 article-title: An efficient iterative approach to explainable feature learning publication-title: IEEE Transactions on Neural Networks and Learning Systems – volume: 144 start-page: 603 year: 2021 ident: 10.1016/j.neunet.2024.106473_b17 article-title: Natural and Artificial Intelligence: A brief introduction to the interplay between AI and neuroscience research publication-title: Neural Networks doi: 10.1016/j.neunet.2021.09.018 – volume: 513 start-page: 165 year: 2022 ident: 10.1016/j.neunet.2024.106473_b23 article-title: Explaining deep neural networks: A survey on the global interpretation methods publication-title: Neurocomputing doi: 10.1016/j.neucom.2022.09.129 – ident: 10.1016/j.neunet.2024.106473_b41 doi: 10.1007/978-3-031-19775-8_27 – ident: 10.1016/j.neunet.2024.106473_b45 doi: 10.1109/ICCV.2017.444 – year: 2018 ident: 10.1016/j.neunet.2024.106473_b19 – start-page: 839 year: 2018 ident: 10.1016/j.neunet.2024.106473_b2 article-title: Grad-CAM++: Generalized gradient-based visual explanations for deep convolutional networks – start-page: 1 year: 2021 ident: 10.1016/j.neunet.2024.106473_b36 article-title: An efficient iterative approach to explainable feature learning publication-title: IEEE Transactions on Neural Networks and Learning Systems – start-page: 818 year: 2014 ident: 10.1016/j.neunet.2024.106473_b38 article-title: Visualizing and understanding convolutional networks – volume: 158 start-page: 83 year: 2023 ident: 10.1016/j.neunet.2024.106473_b24 article-title: A class of doubly stochastic shift operators for random graph signals and their boundedness publication-title: Neural Networks doi: 10.1016/j.neunet.2022.10.035 – start-page: 8971 year: 2021 ident: 10.1016/j.neunet.2024.106473_b22 article-title: Interpreting and disentangling feature components of various complexity from DNNs – volume: 521 start-page: 27 year: 2023 ident: 10.1016/j.neunet.2024.106473_b15 article-title: BFMNet: Bilateral feature fusion network with multi-scale context aggregation for real-time semantic segmentation publication-title: Neurocomputing doi: 10.1016/j.neucom.2022.11.084 – volume: 500 start-page: 989 year: 2022 ident: 10.1016/j.neunet.2024.106473_b31 article-title: CAMA: Class activation mapping disruptive attack for deep neural networks publication-title: Neurocomputing doi: 10.1016/j.neucom.2022.05.065 – volume: 449 start-page: 443 year: 2021 ident: 10.1016/j.neunet.2024.106473_b35 article-title: A novel domain activation mapping-guided network (DA-GNT) for visual tracking publication-title: Neurocomputing doi: 10.1016/j.neucom.2021.03.056 – start-page: 779 year: 2016 ident: 10.1016/j.neunet.2024.106473_b21 article-title: You only look once: Unified, real-time object detection – year: 2012 ident: 10.1016/j.neunet.2024.106473_b10 article-title: ImageNet classification with deep convolutional neural networks – volume: 155 start-page: 58 year: 2022 ident: 10.1016/j.neunet.2024.106473_b33 article-title: Interpretable artificial intelligence through locality guided neural networks publication-title: Neural Networks doi: 10.1016/j.neunet.2022.08.009 – start-page: 1 year: 2022 ident: 10.1016/j.neunet.2024.106473_b27 article-title: A meta-learning approach for training explainable graph neural networks publication-title: IEEE Transactions on Neural Networks and Learning Systems – volume: 13 start-page: 1772 issue: 9 year: 2021 ident: 10.1016/j.neunet.2024.106473_b6 article-title: Self-matching CAM: A novel accurate visual explanation of CNNs for SAR image interpretation publication-title: Remote Sensing doi: 10.3390/rs13091772 – volume: 13 start-page: 4139 issue: 20 year: 2021 ident: 10.1016/j.neunet.2024.106473_b5 article-title: SC-SM CAM: An efficient visual interpretation of CNN for SAR images target recognition publication-title: Remote Sensing doi: 10.3390/rs13204139 – volume: 31 start-page: 3456 issue: 9 year: 2020 ident: 10.1016/j.neunet.2024.106473_b34 article-title: Extracting relational explanations from deep neural networks: A survey from a neural-symbolic perspective publication-title: IEEE Transactions on Neural Networks and Learning Systems doi: 10.1109/TNNLS.2019.2944672 – ident: 10.1016/j.neunet.2024.106473_b26 – ident: 10.1016/j.neunet.2024.106473_b32 doi: 10.1109/CVPR.2016.308 – volume: 31 year: 2018 ident: 10.1016/j.neunet.2024.106473_b12 article-title: Symbolic graph reasoning meets convolutions publication-title: Advances in Neural Information Processing Systems – volume: 359 start-page: 185 year: 2019 ident: 10.1016/j.neunet.2024.106473_b40 article-title: Visualizing and understanding of learned compressive sensing with residual network publication-title: Neurocomputing doi: 10.1016/j.neucom.2019.05.043 – volume: 36 start-page: 133 issue: 6 year: 2019 ident: 10.1016/j.neunet.2024.106473_b30 article-title: Understanding the basis of graph signal processing via an intuitive example-driven approach publication-title: IEEE Signal Processing Magazine doi: 10.1109/MSP.2019.2929832 – year: 2018 ident: 10.1016/j.neunet.2024.106473_b42 – ident: 10.1016/j.neunet.2024.106473_b9 doi: 10.1109/ICCV48922.2021.00137 – year: 2019 ident: 10.1016/j.neunet.2024.106473_b18 – ident: 10.1016/j.neunet.2024.106473_b37 doi: 10.1109/CVPRW50498.2020.00020 – year: 2021 ident: 10.1016/j.neunet.2024.106473_b39 – ident: 10.1016/j.neunet.2024.106473_b25 doi: 10.1109/ICCV.2017.74 – volume: 189 year: 2021 ident: 10.1016/j.neunet.2024.106473_b16 article-title: Fast spectral clustering method based on graph similarity matrix completion publication-title: Signal Processing doi: 10.1016/j.sigpro.2021.108301 – ident: 10.1016/j.neunet.2024.106473_b13 doi: 10.1109/CVPR52688.2022.01167 – volume: 510 start-page: 193 year: 2022 ident: 10.1016/j.neunet.2024.106473_b14 article-title: SiSL-Net: Saliency-guided self-supervised learning network for image classification publication-title: Neurocomputing doi: 10.1016/j.neucom.2022.09.029 – ident: 10.1016/j.neunet.2024.106473_b1 doi: 10.1109/ICCV.2019.00980 – volume: 10 start-page: 1 issue: 1 year: 2019 ident: 10.1016/j.neunet.2024.106473_b11 article-title: Unmasking clever hans predictors and assessing what machines really learn publication-title: Nature Communications doi: 10.1038/s41467-019-08987-4 – start-page: 248 year: 2009 ident: 10.1016/j.neunet.2024.106473_b4 article-title: ImageNet: A large-scale hierarchical image database – year: 2022 ident: 10.1016/j.neunet.2024.106473_b44 article-title: A probe-feature for specific emitter identification using axiom-based grad-CAM publication-title: Signal Processing doi: 10.1016/j.sigpro.2022.108685 – volume: 33 start-page: 4991 issue: 9 year: 2022 ident: 10.1016/j.neunet.2024.106473_b3 article-title: Semisupervised semantic segmentation by improving prediction confidence publication-title: IEEE Transactions on Neural Networks and Learning Systems doi: 10.1109/TNNLS.2021.3066850 – ident: 10.1016/j.neunet.2024.106473_b7 – ident: 10.1016/j.neunet.2024.106473_b43 doi: 10.1109/CVPR.2016.319 – start-page: 770 year: 2016 ident: 10.1016/j.neunet.2024.106473_b8 article-title: Deep residual learning for image recognition – ident: 10.1016/j.neunet.2024.106473_b28 doi: 10.1109/CVPR46437.2021.01625 – ident: 10.1016/j.neunet.2024.106473_b20 – volume: 34 start-page: 176 issue: 4 year: 2017 ident: 10.1016/j.neunet.2024.106473_b29 article-title: Vertex-frequency analysis: A way to localize graph spectral components [lecture notes] publication-title: IEEE Signal Processing Magazine doi: 10.1109/MSP.2017.2696572 |
| SSID | ssj0006843 |
| Score | 2.46708 |
| Snippet | Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently,... |
| SourceID | proquest pubmed crossref elsevier |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 106473 |
| SubjectTerms | Class activation mapping Clustering algorithm Explainable artificial intelligence Image classification |
| Title | Cluster-CAM: Cluster-weighted visual interpretation of CNNs’ decision in image classification |
| URI | https://dx.doi.org/10.1016/j.neunet.2024.106473 https://www.ncbi.nlm.nih.gov/pubmed/38941740 https://www.proquest.com/docview/3073652846 |
| Volume | 178 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVESC databaseName: ScienceDirect database customDbUrl: eissn: 1879-2782 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0006843 issn: 0893-6080 databaseCode: AIEXJ dateStart: 19950101 isFulltext: true titleUrlDefault: https://www.sciencedirect.com providerName: Elsevier |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3Nb9MwFLe6jQMXvj_Kx2QkBidXaeLUDreqdAI0CodOipBQlDg2pJSkrE0ZN_50nmM7pbBp48Alai27Sfx-ff49-30g9DSjfa6kDEiYhj6h1JckDaKMiCz3-koFKgryptgEm0x4HEfvO52fLhZmPWdlyU9Po8V_FTW0gbB16Ow_iLv9UWiAzyB0uILY4XopwY_mtU5-QEbDt9rcd1-_N3ugQC_XxbJuMm1seRtqr4zJpHGpyG3ZHb0VUnzVPj1CU2ztU7QR48xlfWrSdpTGmXwTSyKNCvnwWZYLaRdH7adjamRX5aes2LS-TL9U66Khtcy48s-rg1F4MGzhMKqbgXHxI7XD7EaFT1uXN6dbOYuIz0ytoZ48o80pZFPUx6rUvg6HDc7U9mbjYdYrZQ3v2dM37W26byfXnrxLDo-PjpLpOJ4-W3wjuu6YPp-3RVh20J7Pwgj04t7w9Th-067mA26CNNyDuvDLxkfw7xufR2_OM18aGjO9ga5Z-wMPDW5uoo4sb6HrrrYHtqr-Nvr4G4xe4D9BhA2I8DaIcKWwBtFz7CAEHXADIbwNoTvo-HA8Hb0ithQHET5jK9JnXubnDNi3zJQ-3PfyoJ9l1JNhKEORM3h3wYBv5VSFXAQ5BbMelgvOhCczKoK7aLesSnkfYQWEe8A1M889KqXiUapEFlDGI-DuKu2iwM1gImyeel0uZZ44h8RZYuY90fOemHnvItKOWpg8LRf0Z044ieWahkMmAK4LRj5xskxAFevztbSUVb1M9HI5CIHvDbronhFy-yxgF1Aw_r0Hlxj9EF3d_H8eod3VSS0foytivSqWJ_toh8V838L0F1C6s38 |
| linkProvider | Elsevier |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Cluster-CAM%3A+Cluster-weighted+visual+interpretation+of+CNNs%27+decision+in+image+classification&rft.jtitle=Neural+networks&rft.au=Feng%2C+Zhenpeng&rft.au=Ji%2C+Hongbing&rft.au=Dakovi%C4%87%2C+Milo%C5%A1&rft.au=Cui%2C+Xiyang&rft.date=2024-10-01&rft.issn=1879-2782&rft.eissn=1879-2782&rft.volume=178&rft.spage=106473&rft_id=info:doi/10.1016%2Fj.neunet.2024.106473&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0893-6080&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0893-6080&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0893-6080&client=summon |