{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T13:24:18Z","timestamp":1787059458569,"version":"build-2736575974"},"reference-count":29,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T00:00:00Z","timestamp":1787011200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Deep convolutional neural networks learn rich feature representations; however, the final classification head may not fully align with this feature space after end-to-end training, leading to underutilization of discriminative information. To address this limitation, we propose a lightweight post-hoc redundant head mechanism that improves feature-space adaptation without modifying or retraining the backbone network.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>The proposed approach introduces an additional shallow classification head trained from scratch on frozen feature representations augmented with frozen class logits. This design preserves training stability while enabling broader learning-rate exploration within an expanded signal space. The inclusion of frozen logits provides class-aware priors that stabilize optimization and support improved adaptation within the frozen feature manifold, allowing recovery of residual discriminative information with negligible computational overhead. The method was evaluated using a DenseNet-121 backbone enhanced with a Group\u2013Squeeze-Excitation\u2013Shuffle (GSSE) feature block on the HAM10000 skin lesion dataset through seven-class training and clinically relevant binary evaluation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>The proposed method improved binary classification accuracy from *91.06% to 92.42%* (+1.36 percentage points) and macro-averaged F1-score from *0.8588 to 0.8748* (+0.0160). It also outperformed the strongest baseline accuracy of *91.11%* by *1.31 percentage points*, while introducing only negligible computational overhead.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These findings demonstrate that lightweight post-hoc redundant head learning with logit-aware feature augmentation provides a stable, computationally efficient, and practical strategy for enhancing feature-space adaptation and improving medical image classification performance without requiring modification or retraining of the backbone network.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2026.1867177","type":"journal-article","created":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T12:27:23Z","timestamp":1787056043000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["GSARC: a group shuffle activated redundant connection framework for post-hoc feature space adaptation in skin lesion classification on the HAM10000 dataset"],"prefix":"10.3389","volume":"9","author":[{"given":"Kumar","family":"Abhishek","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering (SCOPE), Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Muralitharan Aishwaryaa","family":"Shree","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering (SCOPE), Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vinesh Kannaa","family":"Balaji","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering (SCOPE), Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rathna","family":"R","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering (SCOPE), Vellore Institute of Technology","place":["Chennai, India"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,18]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"105306","DOI":"10.1016\/j.bspc.2023.105306","article-title":"Skin cancer diagnosis: leveraging deep hidden features and ensemble classifiers for early detection and classification","volume":"88","author":"Akilandasowmya","year":"2024","journal-title":"Biomed. Signal Process. Control"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"100036","DOI":"10.1016\/j.mlwa.2021.100036","article-title":"An enhanced technique of skin cancer classification using deep convolutional neural network with transfer learning models","volume":"5","author":"Ali","year":"2021","journal-title":"Mach. Learn. Appl."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"100034","DOI":"10.1016\/j.neuri.2021.100034","article-title":"Multiclass skin cancer classification using efficientnets\u2013a first step towards preventing skin cancer","volume":"2","author":"Ali","year":"2022","journal-title":"Neurosci. Inform."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"801","DOI":"10.31661\/jbpe.v0i0.2004-1107","article-title":"A deep learning approach to skin cancer detection in dermoscopy images","volume":"10","author":"Ameri","year":"2020","journal-title":"J. Biomed. Phys. Eng."},{"key":"ref5","doi-asserted-by":"publisher","first-page":"97","DOI":"10.3390\/bioengineering9030097","article-title":"Machine learning and deep learning algorithms for skin cancer classification from dermoscopic images","volume":"9","author":"Bechelli","year":"2022","journal-title":"Bioengineering"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"100036","DOI":"10.1016\/j.cmpbup.2021.100036","article-title":"Bilsk: a bilinear convolutional neural network approach for skin lesion classification","volume":"1","author":"Calder\u00f3n","year":"2021","journal-title":"Comput. Methods Programs Biomed. Update"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"28477","DOI":"10.1007\/s11042-020-09388-2","article-title":"A multi-class skin cancer classification using deep convolutional neural networks","volume":"79","author":"Chaturvedi","year":"2020","journal-title":"Multimed. Tools Appl."},{"key":"ref8","doi-asserted-by":"publisher","first-page":"4963","DOI":"10.3390\/s22134963","article-title":"On the automatic detection and classification of skin cancer using deep transfer learning","volume":"22","author":"Fraiwan","year":"2022","journal-title":"Sensors"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1155\/2024\/3022192","article-title":"Skin cancer segmentation and classification using vision transformer for automatic analysis in dermatoscopy-based noninvasive digital system","volume":"2024","author":"Himel","year":"2024","journal-title":"Int. J. Biomed. Imaging"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"310","DOI":"10.1111\/1346-8138.15683","article-title":"Development of a light-weight deep learning model for cloud applications and remote diagnosis of skin cancers","volume":"48","author":"Huang","year":"2021","journal-title":"J. Dermatol."},{"key":"ref11","doi-asserted-by":"publisher","first-page":"8142","DOI":"10.3390\/s21238142","article-title":"Deep learning-based transfer learning for classification of skin cancer","volume":"21","author":"Jain","year":"2021","journal-title":"Sensors"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"e0293266","DOI":"10.1371\/journal.pone.0293266","article-title":"Effnet: a skin cancer classification model based on feature fusion and random forests","volume":"18","author":"Ma","year":"2023","journal-title":"PLoS One"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"3604","DOI":"10.3390\/cancers15143604","article-title":"SBXception: a shallower and broader Xception architecture for efficient classification of skin lesions","volume":"15","author":"Mehmood","year":"2023","journal-title":"Cancer"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"559","DOI":"10.31661\/jbpe.v0i0.2207-1517","article-title":"Skin cancer detection based on deep learning","volume":"12","author":"Mehr","year":"2022","journal-title":"J. Biomed. Phys. Eng."},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1241","DOI":"10.3390\/jcm8081241","article-title":"Risk-aware machine learning classifier for skin lesion diagnosis","volume":"8","author":"Mobiny","year":"2019","journal-title":"J. Clin. Med."},{"key":"ref16","doi-asserted-by":"publisher","first-page":"100659","DOI":"10.1016\/j.imu.2021.100659","article-title":"An approach for multiclass skin lesion classification based on ensemble learning","volume":"25","author":"Rahman","year":"2021","journal-title":"Inform. Med. Unlocked"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"1338","DOI":"10.3390\/diagnostics14131338","article-title":"Integration of localized, contextual, and hierarchical features in deep learning for improved skin lesion classification","volume":"14","author":"Ramamurthy","year":"2024","journal-title":"Diagnostics"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"14","DOI":"10.21037\/jmai-23-38","article-title":"Deep ensemble learning using a demographic machine learning risk stratifier for binary classification of skin lesions using dermatoscopic images","volume":"6","author":"Roge","year":"2023","journal-title":"J. Med. Artif. Intell."},{"key":"ref19","doi-asserted-by":"publisher","first-page":"26795","DOI":"10.1007\/s11042-021-11000-0","article-title":"Deep learning design for benign and malignant classification of skin lesions: a new approach","volume":"80","author":"Salamaa","year":"2021","journal-title":"Multimed. Tools Appl."},{"key":"ref20","doi-asserted-by":"publisher","first-page":"32643","DOI":"10.1007\/s11042-022-13081-x","article-title":"Automated deep learning approach for classification of malignant melanoma and benign skin lesions","volume":"81","author":"Salma","year":"2022","journal-title":"Multimed. Tools Appl."},{"key":"ref21","doi-asserted-by":"publisher","first-page":"18134","DOI":"10.1038\/s41598-022-22644-9","article-title":"Skin lesion classification of dermoscopic images using machine learning and convolutional neural network","volume":"12","author":"Shetty","year":"2022","journal-title":"Sci. Rep."},{"key":"ref22","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","article-title":"A survey on image data augmentation for deep learning","volume":"6","author":"Shorten","year":"2019","journal-title":"J. Big Data"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"e0276836","DOI":"10.1371\/journal.pone.0276836","article-title":"Skinet: a deep learning framework for skin lesion diagnosis with uncertainty estimation and explainability","volume":"17","author":"Singh","year":"2022","journal-title":"PLoS One"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"2852","DOI":"10.3390\/s21082852","article-title":"Classification of skin disease using deep learning neural networks with MobileNet V2 and LSTM","volume":"21","author":"Srinivasu","year":"2021","journal-title":"Sensors"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"e0284437","DOI":"10.1371\/journal.pone.0284437","article-title":"Design and validation of a new machine-learning-based diagnostic tool for the differentiation of dermatoscopic skin cancer images","volume":"18","author":"Tajerian","year":"2023","journal-title":"PLoS One"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"180161","DOI":"10.1038\/sdata.2018.161","article-title":"The ham10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions","volume":"5","author":"Tschandl","year":"2018","journal-title":"Sci. Data"},{"key":"ref27","first-page":"7794","author":"Wang","year":"2018"},{"key":"ref28","first-page":"6848","author":"Zhang","year":"2018"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"105858","DOI":"10.1016\/j.bspc.2023.105858","article-title":"A deep learning outline aimed at prompt skin cancer detection utilizing gated recurrent unit networks and improved orca predation algorithm","volume":"90","author":"Zhang","year":"2024","journal-title":"Biomed. Signal Process. Control"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1867177\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T12:27:24Z","timestamp":1787056044000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1867177\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,18]]},"references-count":29,"alternative-id":["10.3389\/frai.2026.1867177"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1867177","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,18]]},"article-number":"1867177"}}