ORIGINAL RESEARCH article
Front. Artif. Intell.
Sec. Medicine and Public Health
Volume 8 - 2025 | doi: 10.3389/frai.2025.1583079
Spectral Entropic Radiomics Feature Extraction (SERFE): An Adaptive Approach for Glioblastoma Disease Classification
Provisionally accepted
- 1B M S Institute of Technology and Management, Bangalore, India
- 2Shiv Nadar University, Greater Noida, Uttar Pradesh, India
- 3Itron Inc Bangalore, Bangalore, India
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Radiomics-based glioblastoma classification relies on feature extraction methods that can capture tumor heterogeneity while still being fast and efficient. Widely used tools like PyRadiomics and CaPTk extracts extensive handcrafted features, which can lead to redundancy and the need for further optimization in classification tasks. This study presents Spectral Entropic Radiomics Feature Extraction (SERFE), which combines spectral frequency decomposition, entropy-driven selection, and graph-based encoding to improve the characterization of tumors using radiomics. SERFE breaks down changes in voxel intensity into spectral components, uses entropy weighting to give more weight to useful features, and uses graph encoding to keep the shape of the space. Analysis on the Cancer Imaging Archive (TCIA) dataset show that this method is better than traditional ones, with a classification accuracy of 91.7%. The pipeline initiates with 2,260 features per patient and refines them into a final compact set of 350 features, ensuring a 92% stability score and significantly reduced redundancy. SERFE improves feature extraction methods in neuro-oncology and is made to work well with radiomics-based clinical workflows.
Keywords: Spectral Radiomics, Entropy-Weighted Feature Selection, Graph-Theoretic Encoding, Adaptive Radiomics Modelling, glioblastoma classification, TCIA-Based Radiomics, Feature Redundancy reduction
Received: 25 Feb 2025; Accepted: 18 Jun 2025.
Copyright: © 2025 V L, A, Natarajan and N. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Sowmya V L, B M S Institute of Technology and Management, Bangalore, India
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.