A Hybrid QAOA-Based Feature-Selection Framework with Support Vector Machine for Binary Brain Tumor Detection from MRI Images
Computer and Decision Making: An International Journal · 04.11.2026 · 3 (6S) · 3854-3861
DOI: 10.1922/comdem.v3i6s.18832 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer and Decision Making: An International Journal, online erschienen am 04.11.2026. Verfasst von Para Rajesh und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Brain tumor detection from magnetic resonance imaging (MRI) is a challenging image-analysis task. This study investigates a hybrid classical–quantum pipeline for binary image-level classification on the BRISC dataset. The workflow converts images to grayscale, resizes them to 224 × 224 pixels, applies Gaussian denoising and CLAHE, extracts handcrafted intensity, statistical, and gray-level co-occurrence matrix (GLCM) descriptors, screens candidate features using training-only mutual information, formulates a QUBO objective, and uses a depth-one quantum approximate optimization algorithm (QAOA) to select features for an RBF-kernel support vector machine (SVM). A source-code audit performed for this revision identified an off-diagonal coefficient-scaling error in the QUBO construction. The uploaded code archive did not include the BRISC images, generated feature tables, QAOA outputs, or execution logs; therefore, the corrected pipeline could not be rerun and no corrected performance metrics or selected feature subset can be reported. Previous reported test values are only provisional historical values and are not to be considered as test values of the corrected implementation. As a result, the manuscript does not make any claims about a validated QAOA advantage, but rather provides the methodology and explicit reproducibility status. This is an experimental research prototype and has not yet been validated by the outside or clinically
Zitieren
APA
Rajesh, P., Kumar, M. H., Malleswari, D. N., Sirisha, G., Elamathi, E., & Kumar, A. P. (2026). A Hybrid QAOA-Based Feature-Selection Framework with Support Vector Machine for Binary Brain Tumor Detection from MRI Images. Computer and Decision Making: An International Journal, 3(6S), 3854-3861. https://doi.org/10.1922/comdem.v3i6s.18832
Vancouver
Rajesh P, Kumar MH, Malleswari DN, Sirisha G, Elamathi E, Kumar AP. A Hybrid QAOA-Based Feature-Selection Framework with Support Vector Machine for Binary Brain Tumor Detection from MRI Images. Computer and Decision Making: An International Journal. 2026;3(6S):3854-3861. doi:10.1922/comdem.v3i6s.18832
BibTeX
@article{Rajesh2026,
author = {Rajesh, Para and Kumar, Mannepalli Hemanth and Malleswari, D Naga and Sirisha, G.N.V.G and Elamathi, E. and Kumar, Amaravarapu Pramod},
title = {A Hybrid QAOA-Based Feature-Selection Framework with Support Vector Machine for Binary Brain Tumor Detection from MRI Images},
journal = {Computer and Decision Making: An International Journal},
year = {2026},
volume = {3},
number = {6S},
pages = {3854--3861},
doi = {10.1922/comdem.v3i6s.18832}
}
EUVIMED