A Unified Tri-Ensemble Framework with Consensus Explainability (TXCE) for Multiclass Brain Tumor Classification from MRI
Main Article Content
Abstract
This paper proposes TXCE (Tri-Ensemble with Consensus Explainability), a unified architecture for multiclass brain tumor classification from MRI that jointly targets three properties existing systems have so far achieved only in isolation: ensemble robustness, fine-grained subtype discrimination, and multi-method explanation consensus. TXCE combines a performance-weighted soft-voting ensemble of heterogeneous CNN backbones (EfficientNetB0, DenseNet121, and a noise-augmented dual-input branch) with a novel Consensus Explanation Module (CEM) that fuses SHAP, LIME, and Grad-CAM attribution maps into a single composite saliency map together with a quantitative cross-method agreement score. Unlike prior work, which pairs a single classifier design with a fixed explanation technique, TXCE treats explanation fusion itself as a first-class design objective: regions flagged as important by all three independent explanation mechanisms are surfaced as a high-confidence consensus region, giving clinicians an explicit, quantifiable signal of explanation reliability rather than a single, unverified saliency map. We formalize the weighted ensemble aggregation rule, the consensus fusion operator, and the agreement metric mathematically; describe a complete training, inference, and edge-cloud deployment protocol; and present the architecture's design rationale against quantitative baselines reported in the literature for binary ensemble detection, single-backbone multiclass classification, and dual-input noise-robust classification. We further report illustrative case studies, built from representative MRI slices, that demonstrate how the constituent SHAP, LIME, and Grad-CAM outputs combine under the proposed fusion operator. The result is a concrete, mathematically specified architecture intended to serve as a direct blueprint for empirical training and clinical validation, together with an analysis of its expected accuracy-robustness-interpretability trade-offs relative to existing single-technique designs
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
References
A. Amin, K. Hasan, S. Zein-Sabatto, D. Chimba, I. Ahmed, and T. Islam, "An explainable AI framework for artificial intelligence of medical things," arXiv preprint arXiv:2403.04130, 2024.
S. Ahmed, S. N. Nobel, and O. Ullah, "An effective deep CNN model for multiclass brain tumor detection using MRI images and SHAP explainability," in Proc. 2023 Int. Conf. on Electrical, Computer and Communication Engineering (ECCE), IEEE, 2023.
L. Gaur, M. Bhandari, T. Razdan, S. Mallik, and Z. Zhao, "Explanation-driven deep learning model for prediction of brain tumour status using MRI image data," Frontiers in Genetics, vol. 13, p. 822666, 2022.
M. T. Ribeiro, S. Singh, and C. Guestrin, "'Why should I trust you?' Explaining the predictions of any classifier," in Proc. 22nd ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, pp. 1135–1144, 2016.
S. M. Lundberg and S.-I. Lee, "A unified approach to interpreting model predictions," Advances in Neural Information Processing Systems, vol. 30, 2017.
R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, "Grad-CAM: Visual explanations from deep networks via gradient-based localization," in Proc. IEEE Int. Conf. on Computer Vision, pp. 618–626, 2017.
V. Petsiuk, A. Das, and K. Saenko, "RISE: Randomized input sampling for explanation of black-box models," in Proc. British Machine Vision Conf. (BMVC), 2018.
W. Samek, A. Binder, G. Montavon, S. Lapuschkin, and K.-R. Müller, "Evaluating the visualization of what a deep neural network has learned," IEEE Trans. Neural Networks and Learning Systems, vol. 28, no. 11, pp. 2660–2673, 2017.
K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proc. IEEE Conf. on Computer Vision and Pattern Recognition, pp. 770–778, 2016.
G. Huang, Z. Liu, L. van der Maaten, and K. Q. Weinberger, "Densely connected convolutional networks," in Proc. IEEE Conf. on Computer Vision and Pattern Recognition, pp. 4700–4708, 2017.
C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, "Rethinking the inception architecture for computer vision," in Proc. IEEE Conf. on Computer Vision and Pattern Recognition, pp. 2818–2826, 2016.
M. Tan and Q. Le, "EfficientNet: Rethinking model scaling for convolutional neural networks," in Proc. Int. Conf. on Machine Learning (ICML), pp. 6105–6114, 2019.
S. Bhuvaji, "Brain tumor classification (MRI)," Kaggle Dataset, 2020. doi:10.34740/kaggle/dsv/1183165.
A. Hamada, "Br35H: Brain tumor detection 2020," Kaggle Dataset, 2021.
H. Sung, J. Ferlay, R. L. Siegel, M. Laversanne, I. Soerjomataram, A. Jemal, and F. Bray, "Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries," CA: A Cancer Journal for Clinicians, vol. 71, no. 3, pp. 209–249, 2021.
J. Kittler, M. Hatef, R. P. W. Duin, and J. Matas, "On combining classifiers," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 20, no. 3, pp. 226–239, 1998