International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | Cross-Domain Semantic Segmentation with Multi-Level Feature Alignment |
|---|---|
| ABSTRACT | Semantic segmentation plays a pivotal role in computer vision, yet its performance severely degrades when applied to unseen domains due to domain shift. Cross-domain semantic segmentation addresses this challenge by transferring knowledge from labeled source domains to unlabeled target domains. This paper presents a comprehensive framework for cross-domain semantic segmentation through multi-level feature alignment (MLFA). Our approach simultaneously aligns features at pixel-level, feature-level, and semantic-level across domains, leveraging both adversarial learning and self-training strategies. We propose a novel multi-level alignment network that incorporates domain-invariant feature extraction, multi-scale feature fusion, and entropy-based uncertainty estimation. Extensive experiments on standard benchmarks demonstrate that our method achieves state-of-the-art performance, significantly reducing the domain gap and improving segmentation accuracy in target domains. The proposed framework shows remarkable generalization capabilities across various domain shifts including synthetic-to-real, daytime-to-nighttime, and cross-city scenarios. |
| TITLE |
|
| AUTHOR | HOJAMYRADOVA MEYREM, CHOUKPIN ADOTO MIGNONKOUN SOUROU YANNICK MSc Student, Dept. of Information Science and Technology, Nanjing Forestry University, Nanjing, China MSc Student, Dept. of CST and Software, Nanjing University of Information Science and Technology, Nanjing, China |
| VOLUME | 187 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1408001 |
| pdf/1_Cross-Domain Semantic Segmentation with Multi-Level Feature Alignment.pdf | |
| KEYWORDS | |
| References |
[1] H. Li, Y. Wang, and Y. Qiang, “A Semi-Supervised Domain Adaptive Medical Image Segmentation Method Based on Dual-Level Multi-Scale Alignment,” Scientific Reports, vol. 15, no. 1, Mar. 2025, doi: 10.1038/s41598-025-93824-6. [2] J. Liu et al., “Unsupervised Domain Adaptation Multi-Level Adversarial Learning-Based Crossing-Domain Retinal Vessel Segmentation,” Computers in Biology and Medicine, vol. 178, p. 108759, Aug. 2024, doi: 10.1016/j.compbiomed.2024.108759. [3] X. Zhang, Y. Chen, Z. Shen, Y. Shen, H. Zhang, and Y. Zhang, “Confidence-and-Refinement Adaptation Model for Cross-Domain Semantic Segmentation,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 7, pp. 9529–9542, Jul. 2022, doi: 10.1109/tits.2022.3140481. [4] J. Huang, D. Guan, A. Xiao, and S. Lu, “Multi-Level Adversarial Network for Domain Adaptive Semantic Segmentation,” Pattern Recognition, vol. 123, p. 108384, Mar. 2022, doi: 10.1016/j.patcog.2021.108384. [5] Z. Xi et al., “A Multilevel-Guided Curriculum Domain Adaptation Approach to Semantic Segmentation for High-Resolution Remote Sensing Images,” IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–17, 2023, doi: 10.1109/tgrs.2023.3281420. [6] L. Wen, Y. Xu, Z. Feng, J. Zhou, L. Zhou, and Y. Wang, “Semi-Supervised Domain Adaptation for Semantic Segmentation via Active Learning With Feature- and Semantic-Level Alignments,” IEEE Transactions on Intelligent Vehicles, pp. 1–11, 2024, doi: 10.1109/tiv.2024.3410368. [7] P. Liu, Y. Ge, L. Duan, W. Li, and F. Lv, “CAFA: Cross-Modal Attentive Feature Alignment for Cross-Domain Urban Scene Segmentation,” IEEE Transactions on Industrial Informatics, vol. 20, no. 10, pp. 11666–11675, Oct. 2024, doi: 10.1109/tii.2024.3412006. [8] D. Peng, Y. Lei, W. Li, P. Zhang, and Y. Guo, “Sparse-to-Dense Feature Matching: Intra and Inter Domain Cross-Modal Learning in Domain Adaptation for 3D Semantic Segmentation,” Thecvf.com, pp. 7108–7117, 2021, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content/ICCV2021/html/Peng_Sparse-to-Dense_Feature_Matching_Intra_and_Inter_Domain_Cross-Modal_Learning_in_ICCV_2021_paper.html [9] B. Zhang, S. Zhao, and R. Zhang, “Cross-Domain Semantic Segmentation of Urban Scenes via Multi-Level Feature Alignment,” 2020 25th International Conference on Pattern Recognition (ICPR), pp. 1912–1917, Jan. 2021, doi: 10.1109/icpr48806.2021.9411915. [10] F. Lv, G. Lin, P. Liu, G. Yang, S. J. Pan, and L. Duan, “Weakly-Supervised Cross-Domain Road Scene Segmentation via Multi-Level Curriculum Adaptation,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 9, pp. 3493–3503, Sep. 2021, doi: 10.1109/tcsvt.2020.3040343. [11] X. Yang, H. Zhou, De Cheng, M. Tian, and N. Wang, “Multi-Level Explicit Feature Alignment Network for Unsupervised Domain Adaptive Person Search,” IEEE Transactions on Circuits and Systems for Video Technology, pp. 1–1, 2026, doi: 10.1109/tcsvt.2026.3704656. [12] C. Xu, X. Zheng, and X. Lu, “Multi-Level Alignment Network for Cross-Domain Ship Detection,” Remote Sensing, vol. 14, no. 10, p. 2389, May 2022, doi: 10.3390/rs14102389. [13] J. Dong, Z. Long, X. Mao, C. Lin, Y. He, and S. Ji, “Multi-Level Alignment Network for Domain Adaptive Cross-Modal Retrieval,” Neurocomputing, vol. 440, pp. 207–219, Jun. 2021, doi: 10.1016/j.neucom.2021.01.114. [14] R. Xie, Y. Fei, J. Wang, Y. Wang, and L. Zhang, “Multi-Level Domain Adaptive Learning for Cross-Domain Detection,” Thecvf.com, pp. 0–0, 2019, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content_ICCVW_2019/html/TASK-CV/Xie_Multi-Level_Domain_Adaptive_Learning_for_Cross-Domain_Detection_ICCVW_2019_paper.html [15] S. Lei, X. Zhang, J. He, F. Chen, B. Du, and C.-T. Lu, “Cross-Domain Few-Shot Semantic Segmentation,” Lecture Notes in Computer Science, pp. 73–90, 2022, doi: 10.1007/978-3-031-20056-4_5. [16] Y. Liu, J. Deng, X. Gao, W. Li, and L. Duan, “BAPA-Net: Boundary Adaptation and Prototype Alignment for Cross-Domain Semantic Segmentation,” Thecvf.com, pp. 8801–8811, 2021, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content/ICCV2021/html/Liu_BAPA-Net_Boundary_Adaptation_and_Prototype_Alignment_for_Cross-Domain_Semantic_Segmentation_ICCV_2021_paper.html [17] J. Luan, J. Ding, and Y. Zhang, “Unsupervised Cross-Domain Radar Target Recognition Using Multilevel Alignment,” IEEE Transactions on Radar Systems, vol. 3, pp. 630–644, 2025, doi: 10.1109/trs.2025.3560355. [18] M. Zhang, X. Zhao, W. Li, Y. Zhang, R. Tao, and Q. Du, “Cross-Scene Joint Classification of Multisource Data With Multilevel Domain Adaption Network,” IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 8, pp. 11514–11526, Aug. 2024, doi: 10.1109/tnnls.2023.3262599. [19] J. Li, F. Ren, M. Liang, J. Yi, and F. Cao, “Cross Domain Object Detection With Multi-Level Domain Feature Refinement,” Computing, vol. 108, no. 2, Feb. 2026, doi: 10.1007/s00607-026-01621-4. [20] Y. Li, Z. Li, A. Su, K. Wang, Z. Wang, and Q. Yu, “Semisupervised Cross-Domain Remote Sensing Scene Classification via Category-Level Feature Alignment Network,” IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–14, 2024, doi: 10.1109/tgrs.2024.3392984. [21] H. Wang, T. Shen, W. Zhang, L.-Y. Duan, and T. Mei, “Classes Matter: A Fine-Grained Adversarial Approach to Cross-Domain Semantic Segmentation,” Lecture Notes in Computer Science, pp. 642–659, 2020, doi: 10.1007/978-3-030-58568-6_38. [22] D. Zhao et al., “Learning Pseudo-Relations for Cross-Domain Semantic Segmentation,” Thecvf.com, pp. 19191–19203, 2023, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content/ICCV2023/html/Zhao_Learning_Pseudo-Relations_for_Cross-domain_Semantic_Segmentation_ICCV_2023_paper.html [23] Fengmao Lv, T. Liang, X. Chen, and G. Lin, “Cross-Domain Semantic Segmentation via Domain-Invariant Interactive Relation Transfer,” Jun. 2020, doi: 10.1109/cvpr42600.2020.00439. [24] G. Rui, M. Danelljan, H. Sun, J. D. Mangas, and Van Gool Luc, “Prompting Diffusion Representations for Cross-Domain Semantic Segmentation,” arXiv.org. Accessed: Jul. 31, 2026. [Online]. Available: https://arxiv.org/abs/2307.02138 [25] X. Luo, W. Chen, Z. Liang, L. Yang, S. Wang, and C. Li, “Crots: Cross-Domain Teacher–Student Learning for Source-Free Domain Adaptive Semantic Segmentation,” International Journal of Computer Vision, vol. 132, no. 1, pp. 20–39, Aug. 2023, doi: 10.1007/s11263-023-01863-1. [26] L. Qing, F. Lv, L. Duan, and B. Gong, “Constructing Self-Motivated Pyramid Curriculums for Cross-Domain Semantic Segmentation: A Non-Adversarial Approach,” Thecvf.com, pp. 6758–6767, 2019, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content_ICCV_2019/html/Lian_Constructing_Self-Motivated_Pyramid_Curriculums_for_Cross-Domain_Semantic_Segmentation_A_Non-Adversarial_ICCV_2019_paper.html [27] Q. Ren, Q. Mao, and S. Lu, “Prototypical Bidirectional Adaptation and Learning for Cross-Domain Semantic Segmentation,” IEEE Transactions on Multimedia, vol. 26, pp. 501–513, 2024, doi: 10.1109/tmm.2023.3266892. [28] Q. Zhou et al., “Uncertainty-Aware Consistency Regularization for Cross-Domain Semantic Segmentation,” Computer Vision and Image Understanding, vol. 221, p. 103448, Aug. 2022, doi: 10.1016/j.cviu.2022.103448. [29] Y. Gao et al., “PyramidCLIP: Hierarchical Feature Alignment for Vision-Language Model Pretraining,” Advances in Neural Information Processing Systems 35, pp. 35959–35970, 2022, doi: 10.52202/068431-2606. [30] H. Hu et al., “Learning Implicit Feature Alignment Function for Semantic Segmentation,” arXiv.org. Accessed: Jul. 31, 2026. [Online]. Available: https://arxiv.org/abs/2206.08655 [31] C. Zhao, B. Qin, S. Feng, W. Zhu, L. Zhang, and J. Ren, “An Unsupervised Domain Adaptation Method Towards Multi-Level Features and Decision Boundaries for Cross-Scene Hyperspectral Image Classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–16, 2022, doi: 10.1109/tgrs.2022.3230378. [32] Y. Liu, J. Wang, C. Huang, Y. Wu, Y. Xu, and X. Cao, “MLFA: Toward Realistic Test Time Adaptive Object Detection by Multi-Level Feature Alignment,” IEEE Transactions on Image Processing, vol. 33, pp. 5837–5848, 2024, doi: 10.1109/tip.2024.3473532. [33] J. Xu, Z. Chen, S. Yang, J. Li, H. Wang, and Edith, “MENTOR: Multi-Level Self-Supervised Learning for Multimodal Recommendation,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, no. 12, pp. 12908–12917, Apr. 2025, doi: 10.1609/aaai.v39i12.33408. [34] X. Pei, Y. Gao, X. Li, L. Gao, and X. Zhao, “Multilevel Feature Alignment Method for Advancing Remaining Useful Life Prediction of Rolling Bearing,” IEEE Internet of Things Journal, vol. 12, no. 12, pp. 22023–22035, Jun. 2025, doi: 10.1109/jiot.2025.3549728. [35] W. Zhu, C. Zhao, S. Feng, and B. Qin, “Multilevel Feature Alignment Based on Spatial Attention Deformable Convolution for Cross-Scene Hyperspectral Image Classification,” IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1–5, 2022, doi: 10.1109/lgrs.2022.3227251. [36] J. Nie, Z. Dong, Z. He, H. Wu, and M. Gao, “FAML-RT: Feature Alignment-Based Multi-Level Similarity Metric Learning Network for a Two-Stage Robust Tracker,” Information Sciences, vol. 632, pp. 529–542, Mar. 2023, doi: 10.1016/j.ins.2023.02.083. [37] W. Zhou, D. Du, L. Zhang, T. Luo, and Y. Wu, “Multi-Granularity Alignment Domain Adaptation for Object Detection,” Thecvf.com, pp. 9581–9590, 2022, Accessed: Jul. 31, 2026. [Online]. Available: https://openaccess.thecvf.com/content/CVPR2022/html/Zhou_Multi-Granularity_Alignment_Domain_Adaptation_for_Object_Detection_CVPR_2022_paper.html |
