{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T04:17:54Z","timestamp":1787026674843,"version":"build-2736575974"},"reference-count":47,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T00:00:00Z","timestamp":1780617600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100017520","name":"Chengdu Medical College","doi-asserted-by":"publisher","award":["24LHSZ1-04"],"award-info":[{"award-number":["24LHSZ1-04"]}],"id":[{"id":"10.13039\/100017520","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:p>Reliable confidence estimation is essential for surgical visual question answering because overconfident errors may compromise clinical decision support in high-risk settings. However, existing methods typically model predictive uncertainty as a single global quantity, which fails to capture the hierarchical uncertainty arising from ambiguous objects, uncertain tool\u2013tissue interactions, and complex scene context. To address this limitation, we propose a scene graph-guided framework that decomposes uncertainty into object-level, relation-level, and scene-level components and adaptively fuses them for confidence-aware surgical visual question answering. The framework further incorporates Dirichlet-based calibration to improve probabilistic quality and support selective answering. Experiments on the Surgical Scene Graph-Question Answering (SSG-QA) benchmark show that the proposed method achieves an overall accuracy of 63.58%, an Expected Calibration Error of 16.97%, and a Risk-Coverage AUC of 0.0861. Our SG-UD framework achieves competitive calibration performance, significantly improving Brier Score and NLL over baselines, while providing more granular uncertainty interpretability. Performance improves across all question types, with the largest gain observed for relation-type questions (+0.92% over MCAN). Ablation studies further show that uncertainty decomposition contributes most to answer accuracy (+0.70%), whereas Dirichlet calibration is critical for improving probabilistic quality. These findings indicate that explicitly modeling uncertainty across semantic levels can enhance the reliability and interpretability of surgical visual question answering and provide fine-grained confidence information for safer artificial intelligence-assisted surgical decision support.<\/jats:p>","DOI":"10.3389\/fmed.2026.1849346","type":"journal-article","created":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T05:01:02Z","timestamp":1780635662000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Scene graph-guided uncertainty decomposition improves confidence calibration in surgical visual question answering"],"prefix":"10.3389","volume":"13","author":[{"given":"Junzhuo","family":"Song","sequence":"first","affiliation":[{"name":"College of Water Resources and Hydropower, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yaoxue","family":"Xu","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zihan","family":"Zhu","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Junshuang","family":"Zhou","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuhaohang","family":"He","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xirui","family":"Chen","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yunsen","family":"Liang","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wei","family":"Chen","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qiurui","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Otorhinolaryngology Head and Neck Surgery, Ya\u2018an People's Hospital","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun","family":"Li","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Sichuan Agricultural University","place":["Yaan, China"]},{"name":"Agricultural Information Engineering Higher Institution Key Laboratory of Sichuan Province","place":["Yaan, China"]},{"name":"Ya\u2018an Digital Agricultural Engineering Technology Research Center","place":["Yaan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,6,5]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3525145","article-title":"Visual question answering in robotic surgery: a comprehensive review","author":"Ding","year":"2025","journal-title":"IEEE Access"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1007\/s11548-024-03141-y","article-title":"Advancing surgical VQA with scene graph knowledge","author":"Yuan","year":"2024"},{"key":"B3","article-title":"On the validity of Bayesian neural networks for uncertainty estimation","author":"Mitros","year":"2019"},{"key":"B4","doi-asserted-by":"publisher","first-page":"7241","DOI":"10.3390\/s21217241","article-title":"Monte Carlo dropout for uncertainty estimation and motor imagery classification","volume":"21","author":"Milan\u00e9s-Hermosilla","year":"2021","journal-title":"Sensors."},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00681","article-title":"Local temperature scaling for probability calibration","author":"Ding","year":"2021","journal-title":"Proceedings of the IEEE International Conference on Computer Vision"},{"key":"B6","doi-asserted-by":"publisher","first-page":"811","DOI":"10.1111\/biom.13612","article-title":"Selective prediction-set models with coverage rate guarantees","volume":"79","author":"Feng","year":"2023","journal-title":"Biometrics."},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02652","article-title":"From pixels to graphs: Open-vocabulary scene graph generation with vision-language models","author":"Li","year":"2024","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2506.02555","article-title":"SurgVLM: a large vision-language model and systematic evaluation benchmark for surgical intelligence","author":"Zeng","year":"2025","journal-title":"arXiv"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.1007\/s13748-025-00404-w","article-title":"Multinomial classification certainty: a new uncertainty metric for multinomial outcome prediction","author":"van Daalen","year":"2025"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2502.14886","article-title":"Surgical scene understanding in the era of foundation AI models: a comprehensive review","author":"Khan","year":"2025","journal-title":"arXiv"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-64610-3_117","article-title":"Surgical tool detection in laparoscopic videos by modeling temporal dependencies between adjacent frames","author":"Jalal","year":"2020","journal-title":"European Medical and Biological Engineering Conference"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2403.05687","article-title":"Scene graph aided radiology report generation","author":"Wang","year":"2024","journal-title":"arXiv"},{"key":"B13","doi-asserted-by":"publisher","first-page":"104083","DOI":"10.1016\/j.media.2026.104083","article-title":"Decoding the surgical scene: a scoping review of scene graphs in surgery","volume":"112","author":"Henriques","year":"2026","journal-title":"Med Image Anal."},{"key":"B14","article-title":"Simple and scalable predictive uncertainty estimation using deep ensembles","author":"Lakshminarayanan","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B15","doi-asserted-by":"publisher","first-page":"701","DOI":"10.1111\/coin.12411","article-title":"Estimating uncertainty in deep learning for reporting confidence to clinicians in medical image segmentation and diseases detection","volume":"37","author":"Ghoshal","year":"2021","journal-title":"Comput Intell."},{"key":"B16","article-title":"Hercules: deep hierarchical attentive multilevel fusion model with uncertainty quantification for medical image classification","author":"Abdar","year":"2022","journal-title":"IEEE Transactions on Medical Imaging"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2403.07185","article-title":"Uncertainty in graph neural networks: a survey","author":"Wang","year":"2024","journal-title":"arXiv"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1007\/s00138-024-01645-w","article-title":"Graph-based relational reasoning network for video question answering","author":"Tan","year":"2025","journal-title":"arXiv"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2406.10251","article-title":"Uncertainty quantification on graph learning: a survey","author":"Chen","year":"2024","journal-title":"arXiv"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.1145\/3394171.3413761","article-title":"Medical visual question answering via conditional reasoning","author":"Zhan","year":"2020","journal-title":"Proceedings of the 28th ACM International Conference on Multimedia"},{"key":"B21","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-032-08009-7_6","article-title":"Kvasir-VQA-x1: a multimodal dataset for medical reasoning and robust MedVQA in gastrointestinal endoscopy","author":"Gautam","year":"2026"},{"key":"B22","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1145\/3689096.3689458","article-title":"A text-image pair GI tract dataset. in","author":"Gautam","year":"2024","journal-title":"Proceedings of the First International Workshop on Vision-Language Models for Biomedical Applications (VLM4Bio'24)"},{"key":"B23","article-title":"A comprehensive survey of scene graphs: Generation and application","author":"Chang","year":"2021","journal-title":"IEEE Transactions on Multimedia"},{"key":"B24","article-title":"What uncertainties do we need in Bayesian deep learning for computer vision?","author":"Kendall","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW50498.2020.00167","article-title":"Evaluating scalable Bayesian deep learning methods for robust computer vision","author":"Gustafsson","year":"2020","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2404.18279","article-title":"Out-of-distribution detection in medical image analysis: a survey","author":"Hong","year":"2024","journal-title":"arXiv"},{"key":"B27","article-title":"On calibration of modern neural networks","author":"Guo","year":"2017","journal-title":"International Conference on Machine Learning"},{"key":"B28","article-title":"Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods","author":"Platt","year":"1999","journal-title":"Advances in Large Margin Classifiers"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.02336","article-title":"Deep deterministic uncertainty: a new simple baseline","author":"Mukhoti","year":"2023","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1109\/TASE.2020.2986503","article-title":"Camera-robot calibration for the da Vinci robotic surgery system","author":"\u00d6zg\u00fcner","year":"2020","journal-title":"IEEE Transactions on Biomedical Engineering"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i7.28543","article-title":"Cross-modal feature distribution calibration for few-shot visual question answering","author":"Zhang","year":"2024","journal-title":"AAAI Conference on Artificial Intelligence"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1141","DOI":"10.1016\/0895-4356(91)90146-Z","article-title":"Assessing predictive accuracy: how to compare brier scores","volume":"44","author":"Redelmeier","year":"1991","journal-title":"J Clin Epidemiol."},{"key":"B33","article-title":"Selective classification for deep neural networks","author":"Geifman","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B34","article-title":"Faster R-CNN: towards real-time object detection with region proposal networks","author":"Ren","year":"2015","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B35","doi-asserted-by":"publisher","DOI":"10.1109\/ICCVW60793.2023.00015","article-title":"Dynamic scene graph representation for surgical video","author":"Holm","year":"2023","journal-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision Workshops"},{"key":"B36","doi-asserted-by":"publisher","first-page":"264","DOI":"10.1109\/TMI.2023.3299518","article-title":"Concept graph neural networks for surgical video understanding","volume":"43","author":"Ban","year":"2023","journal-title":"IEEE Trans Med Imaging."},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.3390\/ai6050101","article-title":"Embedding fear in medical AI: a risk-averse framework for safety and ethics","author":"Thurzo","year":"2025","journal-title":"AI"},{"key":"B38","article-title":"Providing automated real-time technical feedback for virtual reality based surgical training: is the simpler the better?","author":"Wijewickrema","year":"2018","journal-title":"International Conference on Artificial Intelligence in Medicine"},{"key":"B39","doi-asserted-by":"publisher","first-page":"3080","DOI":"10.1007\/s11605-023-05819-1","article-title":"Real-time artificial intelligence navigation-assisted anatomical recognition in laparoscopic colorectal surgery","volume":"27","author":"Ryu","year":"2023","journal-title":"J Gastrointest Surg."},{"key":"B40","doi-asserted-by":"publisher","first-page":"E10","DOI":"10.3171\/2018.7.FOCUS18281","article-title":"Challenges and opportunities for effective data collection in global neurosurgery: traumatic brain injury surveillance experience in Malawi","volume":"45","author":"Quinsey","year":"2018","journal-title":"Neurosurg Focus."},{"key":"B41","doi-asserted-by":"publisher","first-page":"879","DOI":"10.3390\/bioengineering12080879","article-title":"Beyond post hoc explanations: a comprehensive framework for accountable AI in medical imaging through transparency, interpretability, and explainability","volume":"12","author":"Singh","year":"2025","journal-title":"Bioengineering."},{"key":"B42","doi-asserted-by":"publisher","first-page":"S24","DOI":"10.1016\/j.amjmed.2008.02.001","article-title":"Overconfidence in clinical decision making","volume":"121","author":"Croskerry","year":"2008","journal-title":"Am J Med."},{"key":"B43","doi-asserted-by":"publisher","first-page":"1023","DOI":"10.3390\/make5030053","article-title":"Artificial intelligence ethics and challenges in healthcare applications: a comprehensive review in the context of the European GDPR mandate","volume":"5","author":"Amini","year":"2023","journal-title":"MAKE."},{"key":"B44","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1145\/3721201.3721367","article-title":"Ethical AI for healthcare systems: uncertainty-aware, fair federated learning. in","author":"Chen","year":"2025","journal-title":"ACM\/IEEE International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE '25)"},{"key":"B45","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58452-8_13","article-title":"End-to-end object detection with transformers","author":"Carion","year":"2020","journal-title":"European Conference on Computer Vision"},{"key":"B46","doi-asserted-by":"publisher","first-page":"463","DOI":"10.1162\/pres.17.5.463","article-title":"Modeling of tool-tissue interactions for computer-based surgical simulation: a literature review","volume":"17","author":"Misra","year":"2008","journal-title":"Presence."},{"key":"B47","unstructured":"CVQA: culturally-diverse multilingual visual question answering benchmark\n          \n          \n            \n              Romero\n              D\n            \n            \n              Lyu\n              C\n            \n            \n              Wibowo\n              HA\n            \n            \n              Lynn\n              T\n            \n            \n              Hamed\n              I\n            \n            \n              Nanda Kishore\n              A\n            \n          \n          arXiv\n          \n          2024"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1849346\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T05:01:08Z","timestamp":1780635668000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1849346\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,5]]},"references-count":47,"alternative-id":["10.3389\/fmed.2026.1849346"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1849346","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,5]]},"article-number":"1849346"}}