{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T14:44:39Z","timestamp":1785422679113,"version":"3.56.0"},"reference-count":32,"publisher":"MDPI AG","issue":"9","license":[{"start":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T00:00:00Z","timestamp":1777507200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key R&D Program of China","doi-asserted-by":"crossref","award":["2024YFB2505500"],"award-info":[{"award-number":["2024YFB2505500"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Guangxi Science and Technology Major Program","award":["AA24206054"],"award-info":[{"award-number":["AA24206054"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["52372419"],"award-info":[{"award-number":["52372419"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"award":["52372419"],"award-info":[{"award-number":["52372419"]}],"id":[{"id":"https:\/\/ror.org\/01h0zpd94","id-type":"ROR","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Camera-based 3D occupancy prediction commonly relies on bird\u2019s-eye-view (BEV) representations, yet two limitations remain: optimization instability when inserting new modules into pre-trained BEV encoders, and height-agnostic BEV-to-voxel lifting that fails to preserve elevation-aware scene structure. We propose GSH-Occ (Gradient-Shielded and Height-Aware BEV Occupancy Network), a framework that tackles both issues through two complementary mechanisms. Gradient-Shielded Residual Dual Attention (GS-RDA) introduces a zero-initialized residual gate that preserves the identity mapping at initialization, allowing new attention modules to be grafted onto pre-trained encoders without disturbing learned features. Height-Aware Adaptive Lift (HAL) replaces naive channel replication with per-voxel adaptive fusion of BEV features and learnable height embeddings, followed by 3D convolutional refinement to capture vertical structure. On the Occ3D-nuScenes validation benchmark, GSH-Occ achieves 46.92 mIoU, outperforming FlashOcc by +3.40 mIoU. Ablation studies confirm that GS-RDA and HAL target distinct failure modes and yield complementary improvements.<\/jats:p>","DOI":"10.3390\/s26092800","type":"journal-article","created":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T10:47:28Z","timestamp":1777546048000},"page":"2800","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["GSH-Occ: Gradient-Shielded and Height-Aware BEV Occupancy Network"],"prefix":"10.3390","volume":"26","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-0011-5266","authenticated-orcid":false,"given":"Bokai","family":"Ou","sequence":"first","affiliation":[{"name":"Institute of Artificial Intelligence, Xiamen University, Xiamen 361005, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tianhui","family":"Li","sequence":"additional","affiliation":[{"name":"SAIC GM Wuling Automobile Company Co., Ltd., Liuzhou 545007, China"},{"name":"Guangxi Laboratory of New Energy Automobile, Liuzhou 545616, China"},{"name":"Guangxi Key Laboratory of Automobile Four New Features, Liuzhou 545616, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhigui","family":"Lin","sequence":"additional","affiliation":[{"name":"SAIC GM Wuling Automobile Company Co., Ltd., Liuzhou 545007, China"},{"name":"Guangxi Laboratory of New Energy Automobile, Liuzhou 545616, China"},{"name":"Guangxi Key Laboratory of Automobile Four New Features, Liuzhou 545616, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Boao","family":"Wu","sequence":"additional","affiliation":[{"name":"School of Aerospace Engineering, Xiamen University, Xiamen 361102, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pintong","family":"Chen","sequence":"additional","affiliation":[{"name":"Pen-Tung Sah Institute of Micro-Nano Science and Technology, Xiamen University, Xiamen 361005, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhajiacuo","family":"Zhou","sequence":"additional","affiliation":[{"name":"School of Aerospace Engineering, Xiamen University, Xiamen 361102, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yating","family":"Liu","sequence":"additional","affiliation":[{"name":"School of Aerospace Engineering, Xiamen University, Xiamen 361102, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1013-5374","authenticated-orcid":false,"given":"Jingyao","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Aerospace Engineering, Xiamen University, Xiamen 361102, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9123-6817","authenticated-orcid":false,"given":"Jinghua","family":"Guo","sequence":"additional","affiliation":[{"name":"Pen-Tung Sah Institute of Micro-Nano Science and Technology, Xiamen University, Xiamen 361005, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lei","family":"He","sequence":"additional","affiliation":[{"name":"School of Vehicle and Mobility, Tsinghua University, Beijing 100084, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,4,30]]},"reference":[{"key":"ref_1","unstructured":"Huang, J., Huang, G., Zhu, Z., Ye, Y., and Du, D. (2021). Bevdet: High-performance multi-camera 3d object detection in bird-eye-view. arXiv."},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Li, Z., Wang, W., Li, H., Xie, E., Sima, C., Lu, T., Qiao, Y., and Dai, J. (2022). BEVFormer: Learning Bird\u2019s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers. arXiv.","DOI":"10.1007\/978-3-031-20077-9_1"},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Liu, Z., Tang, H., Amini, A., Yang, X., Mao, H., Rus, D.L., and Han, S. (June, January 29). Bevfusion: Multi-task multi-sensor fusion with unified bird\u2019s-eye view representation. Proceedings of the 2023 IEEE International Conference on Robotics and Automation (ICRA), London, UK.","DOI":"10.1109\/ICRA48891.2023.10160968"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Philion, J., and Fidler, S. (2020, January 23\u201328). Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3D. Proceedings of the European Conference on Computer Vision, Glasgow, UK.","DOI":"10.1007\/978-3-030-58568-6_12"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Li, Y., Ge, Z., Yu, G., Yang, J., Wang, Z., Shi, Y., Sun, J., and Li, Z. (2022). BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection. arXiv.","DOI":"10.1609\/aaai.v37i2.25233"},{"key":"ref_6","unstructured":"Yu, Z., Shu, C., Deng, J., Lu, K., Liu, Z., Yu, J., Yang, D., Li, H., and Chen, Y. (2023). Flashocc: Fast and memory-efficient occupancy prediction via channel-to-height plugin. arXiv."},{"key":"ref_7","unstructured":"Huang, J., and Huang, G. (2022). Bevdet4d: Exploit temporal cues in multi-camera 3d object detection. arXiv."},{"key":"ref_8","first-page":"64318","article-title":"Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving","volume":"36","author":"Tian","year":"2023","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Wei, Y., Zhao, L., Zheng, W., Zhu, Z., Zhou, J., and Lu, J. (2023, January 1\u20136). Surroundocc: Multi-camera 3d occupancy prediction for autonomous driving. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Paris, France.","DOI":"10.1109\/ICCV51070.2023.01986"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Cao, A.Q., and De Charette, R. (2022, January 18\u201324). Monoscene: Monocular 3d semantic scene completion. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.00396"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Zhang, Y., Zhu, Z., and Du, D. (2023, January 1\u20136). Occformer: Dual-path transformer for vision-based 3d semantic occupancy prediction. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Paris, France.","DOI":"10.1109\/ICCV51070.2023.00865"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Huang, Y., Zheng, W., Zhang, Y., Zhou, J., and Lu, J. (2023, January 17\u201324). Tri-perspective view for vision-based 3d semantic occupancy prediction. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada.","DOI":"10.1109\/CVPR52729.2023.00890"},{"key":"ref_13","unstructured":"Goyal, P., Doll\u00e1r, P., Girshick, R., Noordhuis, P., Wesolowski, L., Kyrola, A., Tulloch, A., Jia, Y., and He, K. (2017). Accurate, large minibatch sgd: Training imagenet in 1 hour. arXiv."},{"key":"ref_14","unstructured":"Kumar, A., Raghunathan, A., Jones, R., Ma, T., and Liang, P. (2022). Fine-tuning can distort pretrained features and underperform out-of-distribution. arXiv."},{"key":"ref_15","unstructured":"Bachlechner, T., Majumder, B.P., Mao, H.H., Cottrell, G.W., and McAuley, J. (2020). ReZero is All You Need: Fast Convergence at Large Depth. arXiv."},{"key":"ref_16","unstructured":"Houlsby, N., Giurgiu, A., Jastrzebski, S., Morrone, B., De Laroussilhe, Q., Gesmundo, A., Attariyan, M., and Gelly, S. (2019, January 9\u201315). Parameter-efficient transfer learning for NLP. Proceedings of the International Conference on Machine Learning, Long Beach, CA, USA."},{"key":"ref_17","unstructured":"Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., and Chen, W. (2022, January 25\u201329). LoRA: Low-Rank Adaptation of Large Language Models. Proceedings of the International Conference on Learning Representations, Virtual."},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Ma, Q., Tan, X., Qu, Y., Ma, L., Zhang, Z., and Xie, Y. (2024, January 16\u201322). Cotr: Compact occupancy transformer for vision-based 3d occupancy prediction. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.","DOI":"10.1109\/CVPR52733.2024.01884"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Woo, S., Park, J., Lee, J.Y., and Kweon, I.S. (2018, January 8\u201314). Cbam: Convolutional block attention module. Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany.","DOI":"10.1007\/978-3-030-01234-2_1"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S., and Sun, J. (2015, January 7\u201313). Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile.","DOI":"10.1109\/ICCV.2015.123"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Caesar, H., Bankiti, V., Lang, A.H., Vora, S., Liong, V.E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., and Beijbom, O. (2020, January 13\u201319). nuscenes: A multimodal dataset for autonomous driving. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.","DOI":"10.1109\/CVPR42600.2020.01164"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., and Guo, B. (2021, January 11\u201317). Swin transformer: Hierarchical vision transformer using shifted windows. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Montreal, QC, Canada.","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref_23","unstructured":"Loshchilov, I., and Hutter, F. (2017). Decoupled weight decay regularization. arXiv."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"3421","DOI":"10.1109\/TIV.2024.3453293","article-title":"OccFusion: Multi-Sensor Fusion Framework for 3D Semantic Occupancy Prediction","volume":"10","author":"Ming","year":"2024","journal-title":"IEEE Trans. Intell. Veh."},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Li, Y., Yu, Z., Choy, C., Xiao, C., Alvarez, J.M., Fidler, S., Feng, C., and Anandkumar, A. (2023, January 17\u201324). Voxformer: Sparse voxel transformer for camera-based 3d semantic scene completion. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada.","DOI":"10.1109\/CVPR52729.2023.00877"},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Li, Z., Yu, Z., Wang, W., Anandkumar, A., Lu, T., and Alvarez, J.M. (2023, January 1\u20136). Fb-bev: Bev representation from forward-backward view transformations. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Paris, France.","DOI":"10.1109\/ICCV51070.2023.00637"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Wang, Y., Chen, Y., Liao, X., Fan, L., and Zhang, Z. (2024, January 16\u201322). Panoocc: Unified occupancy representation for camera-based 3d panoptic segmentation. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.","DOI":"10.1109\/CVPR52733.2024.01624"},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Hou, J., Li, X., Guan, W., Zhang, G., Feng, D., Du, Y., Xue, X., and Pu, J. (2024, January 13\u201317). Fastocc: Accelerating 3d occupancy prediction by fusing the 2d bird\u2019s-eye view and perspective view. Proceedings of the 2024 IEEE International Conference on Robotics and Automation (ICRA), Yokohama, Japan.","DOI":"10.1109\/ICRA57147.2024.10610625"},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"5613","DOI":"10.1109\/TITS.2025.3539627","article-title":"Geocc: Geometrically enhanced 3d occupancy network with implicit-explicit depth fusion and contextual self-supervision","volume":"26","author":"Tan","year":"2025","journal-title":"IEEE Trans. Intell. Transp. Syst."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"104670","DOI":"10.1016\/j.cviu.2026.104670","article-title":"DepthOcc: Real-time and accurate 3D occupancy prediction via multi-depth fusion and temporal enhancement","volume":"265","author":"Mao","year":"2026","journal-title":"Comput. Vis. Image Underst."},{"key":"ref_31","unstructured":"Chen, C., Wang, Z., Sheng, T., Jiang, Y., Li, Y., Cheng, P., Zhang, L., Chen, K., Hu, Y., and Yang, X. (2025, January 19\u201325). Sa-occ: Satellite-assisted 3d occupancy prediction in real world. Proceedings of the IEEE\/CVF International Conference on Computer Vision, Honolulu, HI, USA."},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Wolters, P., Gilg, J., Teepe, T., Herzog, F., Laouichi, A., Hofmann, M., and Rigoll, G. (2025, January 19\u201323). Unleashing hydra: Hybrid fusion, depth consistency and radar for unified 3d perception. Proceedings of the 2025 IEEE International Conference on Robotics and Automation (ICRA), Atlanta, GA, USA.","DOI":"10.1109\/ICRA55743.2025.11127412"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/9\/2800\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,2]],"date-time":"2026-05-02T04:29:09Z","timestamp":1777696149000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/9\/2800"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,30]]},"references-count":32,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2026,5]]}},"alternative-id":["s26092800"],"URL":"https:\/\/doi.org\/10.3390\/s26092800","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,30]]}}}