{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,16]],"date-time":"2025-12-16T12:51:33Z","timestamp":1765889493074,"version":"build-2065373602"},"reference-count":38,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2025,5,26]],"date-time":"2025-05-26T00:00:00Z","timestamp":1748217600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"German Federal Ministry of Education and Research","award":["01IS22029C"],"award-info":[{"award-number":["01IS22029C"]}]},{"name":"CertML","award":["01IS22029C"],"award-info":[{"award-number":["01IS22029C"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["MAKE"],"abstract":"<jats:p>Autonomous trains require reliable and accurate environmental perception to take over safety-critical tasks from the driver. This paper investigates the application of N-version architectures to rail track detection using Deep Neural Networks (DNNs) as a means to improve the safety of machine learning (ML)-enabled perception systems. We combine three different neural network architectures (WCID, VGG16-UNet, MobileNet\u2013SegNet) in a 3M1I configuration. In this configuration, we apply two fusion methods to increase accuracy and to enable error detection: Maximum Confidence Voting (MCV), combining the DNN predictions at the image level, and Pixel Majority Voting (PMV), a novel approach for combining the predictions at the pixel level. In addition, we implement a new method for evaluating and combining prediction confidence values in the N-version architecture during runtime. We adjust the overall prediction confidence according to the conformity of all individual predictions, which is not possible with an individual network. Our results show that the N-version architecture not only enables a detection of erroneous predictions by utilizing those adjusted confidence values, but it can also partially improve the predictions by using the PMV combination algorithm. This work emphasizes the importance of model diversity and appropriate thresholds for an accurate assessment of prediction safety. These approaches can significantly improve the practical applicability of ML-based systems in safety-critical domains such as rail transportation.<\/jats:p>","DOI":"10.3390\/make7020049","type":"journal-article","created":{"date-parts":[[2025,5,27]],"date-time":"2025-05-27T11:12:57Z","timestamp":1748344377000},"page":"49","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Using N-Version Architectures for Railway Segmentation with Deep Neural Networks"],"prefix":"10.3390","volume":"7","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7129-3339","authenticated-orcid":false,"given":"Philipp","family":"Ja\u00df","sequence":"first","affiliation":[{"name":"HTW Berlin, University of Applied Sciences, Wilhelminenhofstra\u00dfe 75A, 12459 Berlin, Germany"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0461-2058","authenticated-orcid":false,"given":"Carsten","family":"Thomas","sequence":"additional","affiliation":[{"name":"HTW Berlin, University of Applied Sciences, Wilhelminenhofstra\u00dfe 75A, 12459 Berlin, Germany"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2025,5,26]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"362","DOI":"10.1002\/rob.21918","article-title":"A survey of deep learning techniques for autonomous driving","volume":"37","author":"Grigorescu","year":"2020","journal-title":"J. Field Robot."},{"key":"ref_2","unstructured":"Amodei, D., Olah, C., Steinhardt, J., Christiano, P., Schulman, J., and Man\u00e9, D. (2016). Concrete problems in AI safety. arXiv."},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Zhao, X., Banks, A., Sharp, J., Robu, V., Flynn, D., Fisher, M., and Huang, X. (2020, January 15\u201318). A safety framework for critical systems utilising deep neural networks. Proceedings of the International Conference on Computer Safety, Reliability, and Security, Lisbon, Portugal.","DOI":"10.1007\/978-3-030-54549-9_16"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Dmitriev, K., Schumann, J., and Holzapfel, F. (2021, January 3\u20137). Toward Certification of Machine-Learning Systems for Low Criticality Airborne Applications. Proceedings of the 2021 IEEE\/AIAA 40th Digital Avionics Systems Conference (DASC), San Antonio, TX, USA.","DOI":"10.1109\/DASC52595.2021.9594467"},{"key":"ref_5","unstructured":"(2015). Bahnanwendungen\u2013Betriebsleit-und Zugsicherungssysteme f\u00fcr den st\u00e4dtischen schienengebundenen Personennahverkehr\u2014Teil 1: Systemgrunds\u00e4tze und Grundlegende Konzepte (Standard No. DIN EN 62290-1: 2014)."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"91427","DOI":"10.1109\/ACCESS.2021.3091550","article-title":"Deployment of Autonomous Trains in Rail Transportation: Current Trends and Existing Challenges","volume":"9","author":"Singh","year":"2021","journal-title":"IEEE Access"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"799","DOI":"10.1177\/0361198120925074","article-title":"Investigating Market Potentials and Operational Scenarios of Virtual Coupling Railway Signaling","volume":"2674","author":"Aoun","year":"2020","journal-title":"Transp. Res. Rec. J. Transp. Res. Board"},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Trentesaux, D., Dahyot, R., Ouedraogo, A., Arenas, D., Lefebvre, S., Schon, W., Lussier, B., and Cheritel, H. (2018, January 19\u201322). The Autonomous Train. Proceedings of the 2018 13th Annual Conference on System of Systems Engineering (SoSE), Paris, France.","DOI":"10.1109\/SYSOSE.2018.8428771"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"149","DOI":"10.1177\/09544097211041020","article-title":"Use cases for obstacle detection and track intrusion detection systems in the context of new generation of railway traffic management systems","volume":"236","author":"Hyde","year":"2022","journal-title":"Proc. Inst. Mech. Eng. Part F J. Rail Rapid Transit"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Li, H., Zhang, Q., Zhao, D., and Chen, Y. (2020, January 19\u201324). RailNet: An Information Aggregation Network for Rail Track Segmentation. Proceedings of the 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK.","DOI":"10.1109\/IJCNN48605.2020.9206967"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Tao, Z., Ren, S., Shi, Y., Wang, X., and Wang, W. (2021). Accurate and Lightweight RailNet for Real-Time Rail Line Detection. Electronics, 10.","DOI":"10.3390\/electronics10162038"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Belyaev, S., Popov, I., Shubnikov, V., Popov, P., Boltenkova, E., and Savchuk, D. (2020, January 20\u201323). Railroad semantic segmentation on high-resolution images. Proceedings of the 2020 IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC), Rhodes, Greece.","DOI":"10.1109\/ITSC45102.2020.9294722"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Risti\u0107-Durrant, D., Franke, M., and Michels, K. (2021). A Review of Vision-Based On-Board Obstacle Detection and Distance Estimation in Railways. Sensors, 21.","DOI":"10.3390\/s21103452"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"1426","DOI":"10.1109\/TVT.2021.3133327","article-title":"A Topology Guided Method for Rail-Track Detection","volume":"71","author":"Yang","year":"2022","journal-title":"IEEE Trans. Veh. Technol."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"4077","DOI":"10.1109\/TITS.2023.3327996","article-title":"Key Point Estimate Network for Rail-Track Detection","volume":"25","author":"Yang","year":"2024","journal-title":"IEEE Trans. Intell. Transp. Syst."},{"key":"ref_16","unstructured":"Laurent, T. (2024). Train Ego-Path Detection on Railway Tracks Using End-to-End Deep Learning. arXiv."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Zhang, Q., Yan, F., Song, W., Wang, R., and Li, G. (2023). Automatic Obstacle Detection Method for the Train Based on Deep Learning. Sustainability, 15.","DOI":"10.3390\/su15021184"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Drizi, H.K., and Boukadoum, M. (2024, January 19\u201322). CNN Model with Transfer learning and Data Augmentation for Obstacle Detection in Rail Systems. Proceedings of the 2024 IEEE International Symposium on Circuits and Systems (ISCAS), Singapore, Singapore.","DOI":"10.1109\/ISCAS58744.2024.10558372"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Munappy, A., Bosch, J., Olsson, H.H., Arpteg, A., and Brinne, B. (2019, January 28\u201330). Data Management Challenges for Deep Learning. Proceedings of the 2019 45th Euromicro Conference on Software Engineering and Advanced Applications (SEAA), Kallithea-Chalkidiki, Greece.","DOI":"10.1109\/SEAA.2019.00030"},{"key":"ref_20","unstructured":"(2023). Railway Applications\u2014Requirements for Software Development (Standard No. EN 50716:2023). Technical Report."},{"key":"ref_21","unstructured":"Adler, R., Bunte, A., Burton, S., Gro\u00dfmann, J., Jaschke, A., Kleen, P., Lorenz, J.M., Ma, J., Markert, K., and Mee\u00df, H. (2022). Deutsche Normungsroadmap K\u00fcnstliche Intelligenz, DIN e. V. und DKE."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Machida, F. (2019, January 24\u201327). N-Version Machine Learning Models for Safety Critical Systems. Proceedings of the 2019 49th Annual IEEE\/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W), Portland, OR, USA.","DOI":"10.1109\/DSN-W.2019.00017"},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Xu, H., Chen, Z., Wu, W., Jin, Z., Kuo, S.y., and Lyu, M. (2019, January 24\u201327). NV-DNN: Towards Fault-Tolerant DNN Systems with N-Version Programming. Proceedings of the 2019 49th Annual IEEE\/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W), Portland, OR, USA.","DOI":"10.1109\/DSN-W.2019.00016"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Wu, A., Rubaiyat, A.H.M., Anton, C., and Alemzadeh, H. (2018, January 15\u201318). Model Fusion: Weighted N-Version Programming for Resilient Autonomous Vehicle Steering Control. Proceedings of the 2018 IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW), Memphis, TN, USA.","DOI":"10.1109\/ISSREW.2018.00-11"},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Gujarati, A., Gopalakrishnan, S., and Pattabiraman, K. (2020, January 12\u201315). New Wine in an Old Bottle: N-Version Programming for Machine Learning Components. Proceedings of the 2020 IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW), Coimbra, Portugal.","DOI":"10.1109\/ISSREW51248.2020.00086"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"483","DOI":"10.1007\/s11623-023-1803-z","article-title":"CertML: Initial Steps Towards Using N-Version Neural Networks for Improving AI Safety","volume":"47","author":"Jass","year":"2023","journal-title":"Datenschutz Und Datensicherheit\u2014DuD"},{"key":"ref_27","first-page":"3523","article-title":"Image Segmentation Using Deep Learning: A Survey","volume":"44","author":"Minaee","year":"2022","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Zendel, O., Murschitz, M., Zeilinger, M., Steininger, D., Abbasi, S., and Beleznai, C. (2019, January 16\u201320). RailSem19: A Dataset for Semantic Rail Scene Understanding. Proceedings of the 2019 IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Long Beach, CA, USA.","DOI":"10.1109\/CVPRW.2019.00161"},{"key":"ref_29","unstructured":"NRK (2025, March 09). Nordlandsbanen: Minute by Minute, Season by Season. Available online: https:\/\/nrkbeta.no\/2013\/01\/15\/nordlandsbanen-minute-by-minute-season-by-season\/."},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Michalke, T., W\u00fcst, C., Feng, D., Dolgov, M., Gl\u00e4ser, C., and Timm, F. (2021, January 19\u201322). Where can i drive? A system approach: Deep ego corridor estimation for robust automated driving. Proceedings of the 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), Indianapolis, IN, USA.","DOI":"10.1109\/ITSC48978.2021.9564647"},{"key":"ref_31","first-page":"234","article-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation","volume":"Volume 9351","author":"Navab","year":"2015","journal-title":"Medical Image Computing and Computer-Assisted Intervention\u2014MICCAI 2015"},{"key":"ref_32","unstructured":"Simonyan, K., and Zisserman, A. (2015). Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv."},{"key":"ref_33","unstructured":"Gupta, D. (2025, February 24). Image Segmentation Keras: Implementation of Segnet, FCN, UNet, PSPNet and Other Models in Keras. Available online: https:\/\/github.com\/divamgupta\/image-segmentation-keras."},{"key":"ref_34","unstructured":"Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., and Adam, H. (2017). MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"2481","DOI":"10.1109\/TPAMI.2016.2644615","article-title":"SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation","volume":"39","author":"Badrinarayanan","year":"2017","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Ziegler, M., Mhasawade, V., K\u00f6ppel, M., Neumaier, P., and Eiselein, V. (2023, January 4\u20137). A Comprehensive Framework for Evaluating Vision-Based on-Board Rail Track Detection. Proceedings of the 2023 IEEE Intelligent Vehicles Symposium (IV), Anchorage, AK, USA.","DOI":"10.1109\/IV55152.2023.10186659"},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Casimiro, A., Ortmeier, F., Schoitsch, E., Bitsch, F., and Ferreira, P. (2020). Safety Concerns and Mitigation Approaches Regarding the Use of Deep Learning in Safety-Critical Perception Tasks. Computer Safety, Reliability, and Security. SAFECOMP 2020 Workshops, Proceedings of the DECSoS 2020, DepDevOps 2020, USDAI 2020, and WAISE 2020, Lisbon, Portugal, 15 September 2020, Springer International Publishing. Lecture Notes in Computer Science.","DOI":"10.1007\/978-3-030-55583-2"},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"Dmitriev, K., Schumann, J., Bostanov, I., Abdelhamid, M., and Holzapfel, F. (2023, January 1\u20135). Runway Sign Classifier: A DAL C Certifiable Machine Learning System. Proceedings of the 2023 IEEE\/AIAA 42nd Digital Avionics Systems Conference (DASC), Barcelona, Spain.","DOI":"10.1109\/DASC58513.2023.10311228"}],"container-title":["Machine Learning and Knowledge Extraction"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2504-4990\/7\/2\/49\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T17:40:51Z","timestamp":1760031651000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2504-4990\/7\/2\/49"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,26]]},"references-count":38,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2025,6]]}},"alternative-id":["make7020049"],"URL":"https:\/\/doi.org\/10.3390\/make7020049","relation":{},"ISSN":["2504-4990"],"issn-type":[{"type":"electronic","value":"2504-4990"}],"subject":[],"published":{"date-parts":[[2025,5,26]]}}}