{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T06:45:11Z","timestamp":1782715511900,"version":"3.54.5"},"reference-count":37,"publisher":"MDPI AG","issue":"13","license":[{"start":{"date-parts":[[2026,6,26]],"date-time":"2026-06-26T00:00:00Z","timestamp":1782432000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Accurate recognition of partially occluded faces remains challenging in unconstrained and real-time environments, especially under masks, partial occlusions, pose variation, and illumination changes. This study presents a controlled comparison of three hybrid multi-face recognition pipelines for robust occluded face recognition. For fair evaluation, all pipelines use the same SCRFD face detector, preprocessing protocol, Linear SVM classifier, and 60% unknown rejection threshold, while varying only the feature extractor: ResNet29, ConvNeXt, and ResNet100 with ArcFace embeddings. To reduce data leakage, models are trained only on normal, non-occluded faces and tested on unseen partially occluded faces. Evaluation is performed on a custom dataset and the public Real-World Occluded Faces dataset, alongside three existing paper methods with publicly available code tested under the same experimental protocol. The SCRFD with ArcFace ResNet100 and Linear SVM pipeline achieved the best results compared to existing papers and our other pipelines, reaching 97.475% real-time accuracy for five faces and over 99% confusion-matrix-based accuracy on the custom dataset. On the ROF dataset, it also achieved closed-set accuracies of 98.66% for sunglasses and 97.92% for masks, with threshold-based accuracies of 96.35% for the sunglass test and 95.14% for the mask test. Furthermore, it obtained EER values below 0.007 and AUC values above 99%. In real-time testing, it achieved 29.25 FPS with 34.18 ms\/frame latency on a GPU-enabled laptop and approximately 5 FPS with 273.4 ms\/frame latency on a Raspberry Pi 4.<\/jats:p>","DOI":"10.3390\/s26134069","type":"journal-article","created":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T05:49:33Z","timestamp":1782712173000},"page":"4069","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Controlled Evaluation of Hybrid Multi-Face Recognition Pipelines for Real-Time Occluded Face Recognition on Edge Devices"],"prefix":"10.3390","volume":"26","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-7782-8135","authenticated-orcid":false,"given":"Shk\u00ebmb","family":"Abdullahu","sequence":"first","affiliation":[{"name":"Department of Computer Science, Faculty of Contemporary Sciences and Technologies, South East European University, 1200 Tetovo, North Macedonia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4922-4753","authenticated-orcid":false,"given":"Arbana","family":"Kadriu","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Faculty of Contemporary Sciences and Technologies, South East European University, 1200 Tetovo, North Macedonia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8545-3740","authenticated-orcid":false,"given":"Marco","family":"Piangerelli","sequence":"additional","affiliation":[{"name":"Computer Science Division, School of Science and Technology, University of Camerino, 62032 Camerino, Italy"},{"name":"R&D Division, Vici & C. S.p.A., 47822 Santarcangelo di Romagna, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,6,26]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Li, S.Z., and Jain, A.K. (2011). Handbook of Face Recognition, Springer. [2nd ed.].","DOI":"10.1007\/978-0-85729-932-1"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Erak\u0131n, M.E., Demir, U., and Ekenel, H.K. (2021, January 15\u201317). On Recognizing Occluded Faces in the Wild. Proceedings of the 2021 International Conference of the Biometrics Special Interest Group (BIOSIG), Virtual.","DOI":"10.1109\/BIOSIG52210.2021.9548293"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"3276704","DOI":"10.1155\/2022\/3276704","article-title":"A Real-Time Framework for Human Face Detection and Recognition in CCTV Images","volume":"2022","author":"Ullah","year":"2022","journal-title":"Math. Probl. Eng."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Basly, H., Ouarda, W., Sayadi, F., Ouni, B., and Alimi, A. (2020). CNN-SVM learning approach based human activity recognition. Proceedings of the International Conference on Image and Signal Processing, Springer International Publishing.","DOI":"10.1007\/978-3-030-51935-3_29"},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"2215","DOI":"10.1007\/s10845-023-02152-x","article-title":"Fine-grained activity classification in assembly based on multi-visual modalities","volume":"35","author":"Chen","year":"2024","journal-title":"J. Intell. Manuf."},{"key":"ref_6","unstructured":"Guo, J., Deng, J., Lattas, A., and Zafeiriou, S. (2022, January 25\u201329). Sample and Computation Redistribution for Efficient Face Detection. Proceedings of the International Conference on Learning Representations (ICLR), Virtual."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Liu, Z., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., and Xie, S. (2022, January 18\u201324). A ConvNet for the 2020s. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Deng, J., Guo, J., Xue, N., and Zafeiriou, S. (2019, January 15\u201320). ArcFace: Additive Angular Margin Loss for Deep Face Recognition. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA.","DOI":"10.1109\/CVPR.2019.00482"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"5305","DOI":"10.1007\/s00500-025-10866-8","article-title":"A deep learning-based novel model for masked face recognition","volume":"29","author":"Kumar","year":"2025","journal-title":"Soft Comput."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"126786","DOI":"10.1016\/j.eswa.2025.126786","article-title":"EfficientFaceV2S: A lightweight model and a benchmarking approach for drone-captured face recognition","volume":"273","author":"Alansari","year":"2025","journal-title":"Expert Syst. Appl."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Li, H., Zhang, Y., Wang, W., Zhang, S., and Zhang, S. (2024). Recovery-Based Occluded Face Recognition by Identity-Guided Inpainting. Sensors, 24.","DOI":"10.3390\/s24020394"},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"112595","DOI":"10.1016\/j.asoc.2024.112595","article-title":"Masked face recognition via dual-branch convolutional self-attention network","volume":"169","author":"Wan","year":"2025","journal-title":"Appl. Soft Comput."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"8878553","DOI":"10.1049\/bme2\/8878553","article-title":"Masked Face Recognition Based on ConvNeXt with Coordinate Attention and AdaCos Loss","volume":"2025","author":"Tang","year":"2025","journal-title":"IET Biom."},{"key":"ref_14","first-page":"867","article-title":"Deep Learning-Based Siamese Neural Network for Masked Face Recognition","volume":"10","author":"Nemer","year":"2025","journal-title":"J. Inf. Syst. Eng. Manag."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"101015","DOI":"10.1016\/j.prime.2025.101015","article-title":"Occluded face recognition using optimum features based on efficient preprocessing and machine learning","volume":"12","author":"Khobragade","year":"2025","journal-title":"e-Prime-Adv. Electr. Eng. Electron. Energy"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Ding, F., Peng, P., Huang, Y., Geng, M., and Tian, Y. (2020, January 12\u201316). Masked Face Recognition with Latent Part Detection. Proceedings of the 28th ACM International Conference on Multimedia, Virtual.","DOI":"10.1145\/3394171.3413731"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"108522","DOI":"10.1016\/j.patcog.2022.108522","article-title":"Learning upper patch attention using dual-branch training strategy for masked face recognition","volume":"126","author":"Zhang","year":"2022","journal-title":"Pattern Recognit."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"105541","DOI":"10.1109\/ACCESS.2025.3576639","article-title":"Hierarchical Multi-Scale Patch Attention and Global Feature-Adaptive Fusion for Robust Occluded Face Recognition","volume":"13","author":"Hejazi","year":"2025","journal-title":"IEEE Access"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"e70256","DOI":"10.1049\/ipr2.70256","article-title":"Enhancing Security and Privacy in Occluded Face Recognition: A Human-Centered GAN-Based Approach for Masked Identities in High-Security Environments","volume":"19","author":"Aslam","year":"2025","journal-title":"IET Image Process."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Fan, Q., Li, Z., Li, J., and Cao, C. (July, January 30). MoDE: Mixture of Diffusion Experts for Any Occluded Face Recognition. Proceedings of the 2025 International Joint Conference on Neural Networks (IJCNN), Rome, Italy.","DOI":"10.1109\/IJCNN64981.2025.11228891"},{"key":"ref_21","first-page":"101893","article-title":"Human face localization and detection in highly occluded unconstrained environments","volume":"61","author":"Alashbi","year":"2025","journal-title":"Eng. Sci. Technol. Int. J."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"33","DOI":"10.1007\/s42452-025-08004-y","article-title":"Real-time face recognition algorithm based on lightweight neural network in the field of computer vision","volume":"8","author":"Ma","year":"2025","journal-title":"Discov. Appl. Sci."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Abdullahu, E., Wache, H., and Piangerelli, M. (2025). Secure and Decentralized Hybrid Multi-Face Recognition for IoT Applications. Sensors, 25.","DOI":"10.3390\/s25185880"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"103350","DOI":"10.1016\/j.asej.2025.103350","article-title":"A novel facial recognition technique with focusing on masked faces","volume":"16","author":"Abdullah","year":"2025","journal-title":"Ain Shams Eng. J."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"28467","DOI":"10.1038\/s41598-025-97490-6","article-title":"Real-time facial recognition via multitask learning on raspberry Pi","volume":"15","author":"Aboluhom","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"He, L., Li, H., Zhang, Q., and Sun, Z. (2018, January 18\u201323). Dynamic Feature Learning for Partial Face Recognition. Proceedings of the 2018 IEEE\/CVF Conference on Computer Vision And Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00737"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Li, C., Ge, S., Zhang, D., and Li, J. (2020, January 12\u201316). Look Through Masks: Towards Masked Face Recognition with De-Occlusion Distillation. Proceedings of the 28th ACM International Conference on Multimedia, Virtual.","DOI":"10.1145\/3394171.3413960"},{"key":"ref_28","unstructured":"Anwar, A., and Raychowdhury, A. (2020). Masked Face Recognition for Secure Authentication. arXiv."},{"key":"ref_29","unstructured":"Wang, Z., Wang, G., Huang, B., Xiong, Z., Hong, Q., Wu, H., Yi, P., Jiang, K., Wang, N., and Pei, Y. (2020). Masked Face Recognition Dataset and Application. arXiv."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"679","DOI":"10.1049\/bme2.12036","article-title":"Occlusion-invariant face recognition using simultaneous segmentation","volume":"10","author":"Zeng","year":"2021","journal-title":"IET Biom."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Huber, M., Boutros, F., Kirchbuchner, F., and Damer, N. (2021). Mask-invariant Face Recognition through Template-level Knowledge Distillation. arXiv.","DOI":"10.1109\/FG52635.2021.9667081"},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Deng, H., Feng, Z., Qian, G., Lv, X., Li, H., and Li, G. (2021). MFCosface: A Masked-Face Recognition Algorithm Based on Large Margin Cosine Loss. Appl. Sci., 11.","DOI":"10.3390\/app11167310"},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Boutros, F., Damer, N., Kirchbuchner, F., and Kuijper, A. (2021). Self-restrained Triplet Loss for Accurate Masked Face Recognition. arXiv.","DOI":"10.1016\/j.patcog.2021.108473"},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Zhao, C., Qin, Y., and Zhang, B. (2023). Adversarially Learning Occlusions by Backpropagation for Face Recognition. Sensors, 23.","DOI":"10.3390\/s23208559"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Zhu, K., He, X., Lv, Z., Zhang, X., Hao, R., He, X., Wang, J., He, J., Zhang, L., and Mu, Z. (2023). A 3D Occlusion Facial Recognition Network Based on a Multi-Feature Combination Threshold. Appl. Sci., 13.","DOI":"10.3390\/app13105950"},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"110574","DOI":"10.1016\/j.patcog.2024.110574","article-title":"Masked face recognition using domain adaptation","volume":"153","author":"Huang","year":"2024","journal-title":"Pattern Recognit."},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Saoud, B., Mohamed, A., Shayea, I., El-Saleh, A., and Alashbi, A. (2025). Review of Masked Face Recognition Based on Deep Learning. Technologies, 13.","DOI":"10.3390\/technologies13070310"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/13\/4069\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T05:54:29Z","timestamp":1782712469000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/13\/4069"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,26]]},"references-count":37,"journal-issue":{"issue":"13","published-online":{"date-parts":[[2026,7]]}},"alternative-id":["s26134069"],"URL":"https:\/\/doi.org\/10.3390\/s26134069","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,26]]}}}