{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,8]],"date-time":"2026-08-08T10:43:23Z","timestamp":1786185803398,"version":"build-2736575974"},"reference-count":29,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,8,8]],"date-time":"2026-08-08T00:00:00Z","timestamp":1786147200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,8,8]],"date-time":"2026-08-08T00:00:00Z","timestamp":1786147200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Universit\u00e4t Augsburg"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    The increasing demand for personalized skincare solutions highlights a significant gap: many consumers struggle to find suitable products without professional guidance. While the commercial potential for tailored product recommendations is vast, a key challenge remains the lack of effective methods for skin profiling via image classification. To address this challenge, this paper introduces a comprehensive benchmark dataset of 3203 standardized facial consumer selfies, annotated across eight primary cosmetic skin features. We establish a transparent baseline evaluation utilizing the open-source medical image classification framework AUCMEDI. A variety of deep learning architectures were trained to recognize and classify various skin features, including sagging skin, wrinkles, under-eye circles, redness, shine, pigment spots, acne, and pore size. The baseline model\u2019s performance was evaluated using the mean absolute error (\n                    <jats:italic>e<\/jats:italic>\n                    ), which appropriately accounts for the ordinal distance in cosmetic grading. Utilizing standard deep learning architectures, the baseline benchmark established promising performance in structural categories like sagging skin (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 1.71$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ) and wrinkles (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 1.40$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ). While achieving satisfactory results for under-eye circles (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 0.32$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), redness (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 2.22$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), and shine (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 0.23$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), the baseline models encountered significant architectural limitations when resolving highly localized or imbalanced features such as acne (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 2.59$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), pore size (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 2.57$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), and pigment spots (\n                    <jats:inline-formula>\n                      <jats:tex-math>$$e = 2.20$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    ), the latter three underperforming a trivial constant-mean predictor.\n                  <\/jats:p>","DOI":"10.1038\/s41598-026-65514-4","type":"journal-article","created":{"date-parts":[[2026,8,8]],"date-time":"2026-08-08T10:04:07Z","timestamp":1786183447000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Do not smile when acquiring consumer selfies for multi-attribute skin profiling and baseline deep learning evaluation"],"prefix":"10.1038","volume":"16","author":[{"given":"Dennis","family":"Hartmann","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dominik","family":"M\u00fcller","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Florian","family":"Auer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriele Marie","family":"Lehner","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Laura","family":"Gockeln","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Anna","family":"Rottenkolber","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriel","family":"Duttler","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Julia","family":"Welzel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Frank","family":"Kramer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,8,8]]},"reference":[{"key":"65514_CR1","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.media.2017.07.005","volume":"42","author":"G Litjens","year":"2017","unstructured":"Litjens, G. et al. A survey on deep learning in medical image analysis. Med. Image Anal. 42, 60\u201388. https:\/\/doi.org\/10.1016\/j.media.2017.07.005 (2017).","journal-title":"Med. Image Anal."},{"key":"65514_CR2","unstructured":"Lee, K., Zung, J., Li, P.\u00a0H., Jain, V. & Seung, H. Superhuman Accuracy on the SNEMI3D Connectomics Challenge (2017)."},{"key":"65514_CR3","doi-asserted-by":"publisher","first-page":"221","DOI":"10.1146\/annurev-bioeng-071516-044442","volume":"19","author":"D Shen","year":"2017","unstructured":"Shen, D., Wu, G. & Suk, H.-I. Deep learning in medical image analysis. Annu. Rev. Biomed. Eng. 19, 221\u2013248. https:\/\/doi.org\/10.1146\/annurev-bioeng-071516-044442 (2017).","journal-title":"Annu. Rev. Biomed. Eng."},{"key":"65514_CR4","doi-asserted-by":"publisher","unstructured":"M\u00fcller, D. et al. DeepGleason: A System for Automated Gleason Grading of Prostate Cancer using Deep Neural Networks. https:\/\/doi.org\/10.48550\/arXiv.2403.16678 (2024).","DOI":"10.48550\/arXiv.2403.16678"},{"key":"65514_CR5","doi-asserted-by":"publisher","first-page":"19683","DOI":"10.1007\/s11042-023-15576-7","volume":"83","author":"R Kumar","year":"2024","unstructured":"Kumar, R., Kumbharkar, P., Vanam, S. & Sharma, S. Medical images classification using deep learning: A survey. Multimed. Tools Appl. 83, 19683\u201319728. https:\/\/doi.org\/10.1007\/s11042-023-15576-7 (2024).","journal-title":"Multimed. Tools Appl."},{"key":"65514_CR6","doi-asserted-by":"publisher","unstructured":"Maier-Hein, L. et al. Metrics reloaded: Recommendations for image analysis validation. https:\/\/doi.org\/10.48550\/arXiv.2206.01653 (2024).","DOI":"10.48550\/arXiv.2206.01653"},{"key":"65514_CR7","doi-asserted-by":"publisher","first-page":"71407","DOI":"10.1109\/ACCESS.2023.3295001","volume":"11","author":"P Vatiwutipong","year":"2023","unstructured":"Vatiwutipong, P., Vachmanus, S., Noraset, T. & Tuarob, S. Artificial intelligence in cosmetic dermatology: A systematic literature review. IEEE Access 11, 71407\u201371425. https:\/\/doi.org\/10.1109\/ACCESS.2023.3295001 (2023) (Conference Name: IEEE Access.).","journal-title":"IEEE Access"},{"key":"65514_CR8","doi-asserted-by":"publisher","unstructured":"M\u00fcller, D. AUCMEDI: A framework for Automated Classification of Medical Images. https:\/\/doi.org\/10.5281\/zenodo.6633540 (2022).","DOI":"10.5281\/zenodo.6633540"},{"key":"65514_CR9","doi-asserted-by":"publisher","unstructured":"Mayer, S., M\u00fcller, D. & Kramer, F. Standardized Medical Image Classification across Medical Disciplines. https:\/\/doi.org\/10.48550\/arXiv.2210.11091 (2022).","DOI":"10.48550\/arXiv.2210.11091"},{"key":"65514_CR10","unstructured":"Gatis, D. rembg Original-date: 2020-08-10T14:38:24Z. (2025)."},{"key":"65514_CR11","doi-asserted-by":"publisher","unstructured":"Huang, G., Liu, Z., Maaten, L. V. D. & Weinberger, K. Q. Densely Connected Convolutional Networks. https:\/\/doi.org\/10.48550\/arXiv.1608.06993 (2018).","DOI":"10.48550\/arXiv.1608.06993"},{"key":"65514_CR12","doi-asserted-by":"publisher","unstructured":"He, K., Zhang, X., Ren, S. & Sun, J. Deep Residual Learning for Image Recognition. https:\/\/doi.org\/10.48550\/arXiv.1512.03385 (2015).","DOI":"10.48550\/arXiv.1512.03385"},{"key":"65514_CR13","doi-asserted-by":"publisher","unstructured":"Xie, S., Girshick, R., Doll\u00e1r, P., Tu, Z. & He, K. Aggregated Residual Transformations for Deep Neural Networks. https:\/\/doi.org\/10.48550\/arXiv.1611.05431 (2017).","DOI":"10.48550\/arXiv.1611.05431"},{"key":"65514_CR14","doi-asserted-by":"publisher","unstructured":"Howard, A. G. et al. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. https:\/\/doi.org\/10.48550\/arXiv.1704.04861 (2017).","DOI":"10.48550\/arXiv.1704.04861"},{"key":"65514_CR15","doi-asserted-by":"publisher","unstructured":"Sandler, M., Howard, A., Zhu, M., Zhmoginov, A. & Chen, L.-C. MobileNetV2: Inverted Residuals and Linear Bottlenecks. https:\/\/doi.org\/10.48550\/arXiv.1801.04381ArXiv:1801.04381 (2019).","DOI":"10.48550\/arXiv.1801.04381"},{"key":"65514_CR16","doi-asserted-by":"publisher","unstructured":"Simonyan, K. & Zisserman, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. https:\/\/doi.org\/10.48550\/arXiv.1409.1556 (2015).","DOI":"10.48550\/arXiv.1409.1556"},{"key":"65514_CR17","doi-asserted-by":"publisher","unstructured":"Liu, Z. et al. A ConvNet for the 2020s. https:\/\/doi.org\/10.48550\/arXiv.2201.03545 (2022).","DOI":"10.48550\/arXiv.2201.03545"},{"key":"65514_CR18","doi-asserted-by":"publisher","unstructured":"Kingma, D. P. & Ba, J. Adam: A Method for Stochastic Optimization. https:\/\/doi.org\/10.48550\/arXiv.1412.6980ArXiv:1412.6980 (2017).","DOI":"10.48550\/arXiv.1412.6980"},{"key":"65514_CR19","doi-asserted-by":"publisher","unstructured":"Deng, J. et al. ImageNet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, 248\u2013255. https:\/\/doi.org\/10.1109\/CVPR.2009.5206848 (2009).","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"65514_CR20","doi-asserted-by":"publisher","unstructured":"Lin, T.-Y., Goyal, P., Girshick, R., He, K. & Doll\u00e1r, P. Focal Loss for Dense Object Detection. https:\/\/doi.org\/10.48550\/arXiv.1708.02002 (2018).","DOI":"10.48550\/arXiv.1708.02002"},{"key":"65514_CR21","doi-asserted-by":"publisher","first-page":"3623","DOI":"10.3390\/electronics11213623","volume":"11","author":"T-R Chang","year":"2022","unstructured":"Chang, T.-R. & Tsai, M.-Y. Classifying conditions of speckle and wrinkle on the human face: A deep learning approach. Electronics 11, 3623. https:\/\/doi.org\/10.3390\/electronics1121362321 (2022).","journal-title":"Electronics"},{"key":"65514_CR22","doi-asserted-by":"publisher","unstructured":"Zhao, T., Zhang, H. & Spoelstra, J. A Computer Vision Application for Assessing Facial Acne Severity from Selfie Images. https:\/\/doi.org\/10.48550\/arXiv.1907.07901 (2019).","DOI":"10.48550\/arXiv.1907.07901"},{"key":"65514_CR23","doi-asserted-by":"publisher","unstructured":"Zheng, Q. et al. Automatic Facial Skin Feature Detection for Everyone. https:\/\/doi.org\/10.48550\/arXiv.2203.16056 (2022).","DOI":"10.48550\/arXiv.2203.16056"},{"key":"65514_CR24","doi-asserted-by":"publisher","first-page":"851","DOI":"10.2147\/CCID.S360450","volume":"15","author":"A Quattrini","year":"2022","unstructured":"Quattrini, A., Bo\u00ebr, C., Leidi, T. & Paydar, R. A deep learning-based facial acne classification system. Clin. Cosmet. Investig. Dermatol. 15, 851\u2013857. https:\/\/doi.org\/10.2147\/CCID.S360450https:\/\/www.tandfonline.com\/doi\/pdf\/10.2147\/CCID.S360450 (2022).","journal-title":"Clin. Cosmet. Investig. Dermatol."},{"key":"65514_CR25","doi-asserted-by":"publisher","DOI":"10.1111\/jocd.70050","volume":"24","author":"RL Draelos","year":"2025","unstructured":"Draelos, R. L., Kesty, C. E. & Kesty, K. R. Artificial intelligence predicts Fitzpatrick skin type, pigmentation, redness, and wrinkle severity from color photographs of the face. J. Cosmet. Dermatol. 24, e70050. https:\/\/doi.org\/10.1111\/jocd.70050https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/jocd.70050 (2025).","journal-title":"J. Cosmet. Dermatol."},{"key":"65514_CR26","doi-asserted-by":"publisher","first-page":"1894","DOI":"10.3390\/diagnostics13111894","volume":"13","author":"H Yoon","year":"2023","unstructured":"Yoon, H., Kim, S., Lee, J. & Yoo, S. Deep-learning-based morphological feature segmentation for facial skin image analysis. Diagnostics 13, 1894. https:\/\/doi.org\/10.3390\/diagnostics13111894 (2023).","journal-title":"Diagnostics"},{"key":"65514_CR27","doi-asserted-by":"publisher","first-page":"989","DOI":"10.3390\/app13020989","volume":"13","author":"M Kim","year":"2023","unstructured":"Kim, M. & Song, M. H. High performing facial skin problem diagnosis with enhanced mask R-CNN and super resolution GAN. Appl. Sci. 13, 989. https:\/\/doi.org\/10.3390\/app13020989 (2023).","journal-title":"Appl. Sci."},{"key":"65514_CR28","doi-asserted-by":"publisher","DOI":"10.1111\/srt.13492","volume":"29","author":"T-H Li","year":"2023","unstructured":"Li, T.-H. et al. Artificial intelligence analysis of over a million Chinese men and women reveals level of dark circle in the facial skin aging process. Skin Res. Technol. 29, e13492. https:\/\/doi.org\/10.1111\/srt.13492 (2023) https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/srt.13492.","journal-title":"Skin Res. Technol."},{"key":"65514_CR29","doi-asserted-by":"publisher","DOI":"10.1111\/srt.13632","volume":"30","author":"GM Lehner","year":"2024","unstructured":"Lehner, G. M. et al. Differences in the annotation between facial images and videos for training an artificial intelligence for skin type determination. Skin Res. Technol. Off. J. Int. Soc. Bioeng. Skin (ISBS) Int. Soc. Digital Imaging Skin (ISDIS) [and] Int. Soc. Skin Imaging (ISSI) 30, e13632. https:\/\/doi.org\/10.1111\/srt.13632 (2024).","journal-title":"Skin Res. Technol. Off. J. Int. Soc. Bioeng. Skin (ISBS) Int. Soc. Digital Imaging Skin (ISDIS) [and] Int. Soc. Skin Imaging (ISSI)"}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-65514-4","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-65514-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-65514-4.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,8]],"date-time":"2026-08-08T10:17:12Z","timestamp":1786184232000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-65514-4"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,8]]},"references-count":29,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["65514"],"URL":"https:\/\/doi.org\/10.1038\/s41598-026-65514-4","relation":{},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,8]]},"assertion":[{"value":"10 February 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"31 July 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"8 August 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Gabriel Duttler and Anna Rottenkolber are employees of GRANDEL the Beautyness Company.","order":1,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"All methods were carried out in accordance with relevant guidelines and regulations. This study was a non-interventional, retrospective analysis of pseudonymized data of healthy subjects, not involving medical drugs or devices. Following an internal review by the University of Augsburg and the University Hospital Augsburg, formal ethical approval was deemed not required for this type of study. A formal data protection agreement was established, and written informed consent was obtained from all participants for the processing and use of their data within the scope of this study.","order":2,"name":"Ethics","label":"Ethical approval and informed consent","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Prof. Dr. Welzel, co-author of this manuscript, consents to the publication of her images (Figs.\n                      \n                      and\n                      \n                      ) in Nature Scientific Reports under the CC-BY-NC-ND license. All other figures and graphics were created by the authors.","order":3,"name":"Ethics","label":"Consent for publication","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"24546"}}