{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,4]],"date-time":"2026-04-04T05:02:47Z","timestamp":1775278967092,"version":"3.50.1"},"reference-count":45,"publisher":"MDPI AG","issue":"7","license":[{"start":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T00:00:00Z","timestamp":1774915200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62472149"],"award-info":[{"award-number":["62472149"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Transformer models have achieved powerful performance in various computer vision tasks. However, their black-box nature severely limits model interpretability and the reliability of real-world applications. Most existing interpretation methods generate explanation maps by perturbing masks from the last layer of the Transformer encoder, but they often overlook uncertain information in masks and detail loss during upsampling and downsampling, resulting in coarse localization, blurred boundaries, and significant background noise in explanations. To address these issues, this paper proposes a self-distillation object segmentation method based on sequential three-way mask and attention fusion (SAF-SD), targeting salient and camouflaged binary object segmentation tasks (sub-tasks of binary pixel-level segmentation). The method consists of two core modules: the sequential three-way mask (S3WM) module and the attention fusion (AF) module. The S3WM module performs strict threshold filtering on masks generated from the final-layer feature maps of the Transformer, aiming to accurately segment foreground objects from backgrounds via binary pixel-level prediction. The AF module aggregates attention matrices across all Transformer encoder layers to construct a cross-layer relation matrix, capturing global semantic dependencies among image patches (e.g., interactions between foreground, background, and edge regions). It then computes the importance score for each patch, refining details and suppressing noise in the initial explanation results. Extensive experimental results demonstrate that SAF-SD significantly outperforms existing baseline methods across key evaluation metrics.<\/jats:p>","DOI":"10.3390\/s26072170","type":"journal-article","created":{"date-parts":[[2026,3,31]],"date-time":"2026-03-31T16:34:08Z","timestamp":1774974848000},"page":"2170","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["SAF-SD: Self-Distillation Object Segmentation Method Based on Sequential Three-Way Mask and Attention Fusion"],"prefix":"10.3390","volume":"26","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-4666-5663","authenticated-orcid":false,"given":"Biao","family":"Wang","sequence":"first","affiliation":[{"name":"School of Information Engineering, Wuhan College, Wuhan 430212, China"},{"name":"Key Laboratory of Green Intelligent Computing Network in Hubei Province, Hubei University of Technology, Wuhan 430068, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jun","family":"Su","sequence":"additional","affiliation":[{"name":"Key Laboratory of Green Intelligent Computing Network in Hubei Province, Hubei University of Technology, Wuhan 430068, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Volodymyr","family":"Kochan","sequence":"additional","affiliation":[{"name":"Department of Information and Computing Systems and Control, West Ukrainian National University, 46009 Ternopil, Ukraine"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2468-3881","authenticated-orcid":false,"given":"Lingyu","family":"Yan","sequence":"additional","affiliation":[{"name":"Key Laboratory of Green Intelligent Computing Network in Hubei Province, Hubei University of Technology, Wuhan 430068, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2026,3,31]]},"reference":[{"key":"ref_1","first-page":"335","article-title":"Occluded object segmentation based on bilayer decoupling strategy and attention mechanism","volume":"45","author":"Lu","year":"2023","journal-title":"J. Electron. Inf. Technol."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"884","DOI":"10.1007\/s11760-025-04512-x","article-title":"Small target detection model based on swin transformer and multiscale fusion","volume":"19","author":"Wang","year":"2025","journal-title":"Signal Image Video Process."},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., and Guo, B. (2021, January 10\u201317). Swin transformer: Hierarchical vision transformer using shifted windows. Proceedings of the 2021 IEEE\/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada.","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"3111","DOI":"10.1109\/TPAMI.2025.3529350","article-title":"Condition-Invariant Semantic Segmentation","volume":"47","author":"Sakaridis","year":"2025","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_5","first-page":"1638","article-title":"A survey of knowledge distillation in deep learning","volume":"45","author":"Shao","year":"2022","journal-title":"Chin. J. Comput."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Wu, D., Chen, P., Yu, X., Li, G., Han, Z., and Jiao, J. (2023, January 1\u20136). Spatial self-distillation for object detection with inaccurate bounding boxes. Proceedings of the 2023 IEEE\/CVF International Conference on Computer Vision (ICCV), Paris, France.","DOI":"10.1109\/ICCV51070.2023.00631"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Lin, S., Xie, H., Wang, B., Yu, K., Chang, X., Liang, X., and Wang, G. (2022, January 18\u201324). Knowledge distillation via the target-aware transformer. Proceedings of the 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.01064"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"20933","DOI":"10.1109\/TITS.2024.3455416","article-title":"TransKD: Transformer knowledge distillation for efficient semantic segmentation","volume":"25","author":"Liu","year":"2024","journal-title":"IEEE Trans. Intell. Transp. Syst."},{"key":"ref_9","unstructured":"Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., and Kaiser, \u0141. (2017, January 4\u20139). Attention is all you need. Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS 2017), Cambridge, MA, USA."},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Wang, X., Wu, H., Wang, L., Chen, J., Li, Y., He, X., Chen, T., Wang, M., and Guo, L. (2025). Enhanced pulmonary nodule detection with U-Net, YOLOv8, and swin transformer. BMC Med. Imaging, 25.","DOI":"10.1186\/s12880-025-01784-0"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Wang, W., Xie, E., Li, X., Fan, D.-P., Song, K., Liang, D., Lu, T., Luo, P., and Shao, L. (2021, January 10\u201317). Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. Proceedings of the 18th IEEE\/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada.","DOI":"10.1109\/ICCV48922.2021.00061"},{"key":"ref_12","first-page":"227","article-title":"Enhanced surface defect detection of cylinder liners using Swin Transformer and YOLOv8","volume":"4","author":"Pan","year":"2025","journal-title":"J. Autom. Intell."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"10212","DOI":"10.1109\/TPAMI.2023.3257878","article-title":"Online knowledge distillation via mutual contrastive learning for visual recognition","volume":"45","author":"Yang","year":"2023","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"e1435","DOI":"10.7717\/peerj-cs.1435","article-title":"A non-negative feedback self-distillation method for salient object detection","volume":"9","author":"Chen","year":"2023","journal-title":"PeerJ Comput. Sci."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"2615","DOI":"10.1007\/s00530-023-01121-x","article-title":"Self-distillation object segmentation via pyramid knowledge representation and transfer","volume":"29","author":"Zheng","year":"2023","journal-title":"Multimed. Syst."},{"key":"ref_16","unstructured":"Smilkov, D., Thorat, N., Kim, B., Vi\u00e9gas, F., and Wattenberg, M. (2017). Smoothgrad: Removing noise by adding noise. arXiv."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Han, S., Ling, Y., Xie, M., Ming, S., and Tang, F. (2025). A Malicious Encrypted Traffic Detection Method Based on Hybrid CNN and Swin Transformer. Proceedings of the International Conference on Cyberspace Simulation and Evaluation (CSE), Springer.","DOI":"10.1007\/978-981-96-4509-1_19"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Binder, A., Montavon, G., Lapuschkin, S., M\u00fcller, K.R., and Samek, W. (2016, January 6\u20139). Layer-wise relevance propagation for neural networks with local renormalization layers. Proceedings of the 25th International Conference on Artificial Neural Networks (ICANN), Barcelona, Spain.","DOI":"10.1007\/978-3-319-44781-0_8"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Chefer, H., Gur, S., and Wolf, L. (2021, January 19\u201325). Transformer interpretability beyond attention visualization. Proceedings of the 34th IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.00084"},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"203","DOI":"10.1007\/s11760-024-03757-2","article-title":"MS-DETR: A real-time multi-scale detection transformer for PCB defect detection","volume":"19","author":"Ji","year":"2025","journal-title":"Signal Image Video Process."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"e70031","DOI":"10.1049\/ipr2.70031","article-title":"Structure-Aware Transformer for Shadow Detection","volume":"19","author":"Sun","year":"2025","journal-title":"IET Image Process."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Chowdhury, A., Paul, D., Mai, Z., Gu, J., Zhang, Z., Mehrab, K.S., Campolongo, E.G., Rubenstein, D., Stewart, C.V., and Karpatne, A. (2025). Prompt-CAM: A Simpler Interpretable Transformer for Fine-Grained Analysis. arXiv.","DOI":"10.1109\/CVPR52734.2025.00413"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"102355","DOI":"10.1016\/j.inffus.2024.102355","article-title":"Deconfounded multi-organ weakly-supervised semantic segmentation via causal intervention","volume":"108","author":"Chen","year":"2024","journal-title":"Inf. Fusion"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"103139","DOI":"10.1016\/j.media.2024.103139","article-title":"Anomaly-guided weakly supervised lesion segmentation on retinal OCT images","volume":"94","author":"Yang","year":"2024","journal-title":"Med. Image Anal."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"497","DOI":"10.26599\/CVM.2025.9450388","article-title":"Point mask transformer for outdoor point cloud semantic segmentation","volume":"11","author":"Li","year":"2025","journal-title":"Comput. Vis. Media"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"1104","DOI":"10.1109\/TNNLS.2023.3326283","article-title":"MGRW-Transformer: Multigranularity random walk Transformer model for interpretable learning","volume":"36","author":"Ding","year":"2023","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"1178","DOI":"10.1007\/s11227-025-07661-5","article-title":"A multilevel integrated supervision self-distillation method","volume":"81","author":"Wang","year":"2025","journal-title":"J. Supercomput."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Xie, W., Li, X., Cao, C., and Zhang, N.L. (2023, January 19\u201325). ViT-CX: Causal explanation of vision transformers. Proceedings of the 32nd International Joint Conference on Artificial Intelligence (IJCAI), San Francisco, CA, USA.","DOI":"10.24963\/ijcai.2023\/174"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Abnar, S., and Zuidema, W. (2020, January 5\u201310). Quantifying attention flow in Transformers. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), Stroudsburg, PA, USA.","DOI":"10.18653\/v1\/2020.acl-main.385"},{"key":"ref_30","first-page":"2737","article-title":"Transformer Interpretation Method Based on Sequential Three-Way Mask and Attention Fusion","volume":"62","author":"Cheng","year":"2025","journal-title":"J. Comput. Res. Dev."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Fan, D., Ji, G., Sun, G., Cheng, M.-M., Shen, J., and Shao, L. (2020, January 13\u201319). Camouflaged object detection. Proceedings of the 2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA.","DOI":"10.1109\/CVPR42600.2020.00285"},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Yang, C., Zhang, L., Lu, H., Ruan, X., and Yang, M.-H. (2013, January 23\u201328). Saliency detection via graph based manifold ranking. Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Portland, OR, USA.","DOI":"10.1109\/CVPR.2013.407"},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Yin, P., Fu, K., and Zhao, Q. (2025). Key Object Detection: Unifying Salient and Camouflaged Object Detection Into One Task. Proceedings of the Chinese Conference on Pattern Recognition and Computer Vision (PRCV), Springer.","DOI":"10.1007\/978-981-97-8858-3_37"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"104789","DOI":"10.1016\/j.dsp.2024.104789","article-title":"DEDBNet: DoG-enhanced dual-branch object detection network for remote sensing object detection","volume":"156","author":"Pan","year":"2025","journal-title":"Digital Signal Process."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"4029","DOI":"10.1109\/TNNLS.2022.3144003","article-title":"SAB Net: A Semantic Attention Boosting Framework for Semantic Segmentation","volume":"36","author":"Ding","year":"2025","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Huang, Z., Wang, X., Huang, L., Huang, C., Wei, Y., and Liu, W. (November, January 27). CCNeT: Criss-cross attention for semantic segmentation. Proceedings of the 2019 IEEE\/CVF International Conference on Computer Vision (ICCV), Seoul, Republic of Korea.","DOI":"10.1109\/ICCV.2019.00069"},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Zhao, X., Pang, Y., Zhang, L., Lu, H., and Zhang, L. (2021, January 23\u201328). Suppress and balance: A simple gated network for salient object detection. Proceedings of the 16th European Conference on Computer Vision (ECCV), Glasgow, UK.","DOI":"10.1007\/978-3-030-58536-5_3"},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"Wu, Z., Su, L., and Huang, Q. (2019, January 15\u201320). Cascaded partial decoder for fast and accurate salient object detection. Proceedings of the 2019 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA.","DOI":"10.1109\/CVPR.2019.00403"},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"728","DOI":"10.1109\/TCSVT.2020.2988768","article-title":"Deep sub-region network for salient object detection","volume":"31","author":"Wang","year":"2022","journal-title":"IEEE Trans. Circuits Syst. Video Technol."},{"key":"ref_40","doi-asserted-by":"crossref","first-page":"3125","DOI":"10.1109\/TIP.2022.3164550","article-title":"EDN: Salient object detection via extremely-downsampled network","volume":"31","author":"Wu","year":"2023","journal-title":"IEEE Trans. Image Process."},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"887","DOI":"10.1109\/TPAMI.2021.3140168","article-title":"PoolNet+: Exploring the potential of pooling for salient object detection","volume":"45","author":"Liu","year":"2024","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_42","doi-asserted-by":"crossref","unstructured":"Hou, Y., Ma, Z., Liu, C., and Loy, C.C. (November, January 27). Learning lightweight lane detection CNNs by self-attention distillation. Proceedings of the 2019 IEEE\/CVF International Conference on Computer Vision (ICCV), Seoul, Republic of Korea.","DOI":"10.1109\/ICCV.2019.00110"},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Zhang, L., Song, J., Gao, A., Chen, J., Bao, C., and Ma, K. (November, January 27). Be your own teacher: Improve the performance of convolutional neural networks via self-distillation. Proceedings of the 2019 IEEE\/CVF International Conference on Computer Vision (ICCV), Seoul, Republic of Korea.","DOI":"10.1109\/ICCV.2019.00381"},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"Sun, D., Yao, A., Zhou, A., and Zhao, H. (2019, January 15\u201320). Deeply-supervised knowledge synergy. Proceedings of the 2019 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA.","DOI":"10.1109\/CVPR.2019.00716"},{"key":"ref_45","doi-asserted-by":"crossref","unstructured":"Li, D., and Chen, Q. (2020, January 13\u201319). Dynamic hierarchical mimicking towards consistent optimization objectives. Proceedings of the 2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA.","DOI":"10.1109\/CVPR42600.2020.00766"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/7\/2170\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,4]],"date-time":"2026-04-04T04:18:50Z","timestamp":1775276330000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/7\/2170"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,31]]},"references-count":45,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2026,4]]}},"alternative-id":["s26072170"],"URL":"https:\/\/doi.org\/10.3390\/s26072170","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,31]]}}}