Multimodal breast cancer diagnosis using feature fusion and deep learning - Takeaways - MDSpire

Integrative Approaches for Breast Cancer Diagnosis through Feature Fusion and Deep Learning Techniques

  • By

  • Varun Malik

  • Tahani Alsubait

  • Mudassir Khan

  • Upasana Lakhina

  • Alaa Menshawi

  • Stuti Mehla

  • Loveleena Mukhija

  • Meteb Altaf

  • Leila Jamel

  • July 21, 2026

Share

  • 1

    Breast cancer diagnosis faces challenges due to the disease's heterogeneity and limitations of traditional imaging modalities.

  • 2

    The proposed multimodal model utilizes attention-based transformers for feature extraction and combines features using the AZO algorithm.

  • 3

    The model achieves high accuracy rates of 98.958%, 97.37%, and 99.438% on MIAS, BreakHis, and combined datasets, respectively.

  • 4

    A lightweight convolutional neural network (LCNN) is employed for classification, ensuring efficiency in resource-limited clinical settings.

  • 5

    The model effectively addresses issues of missing modalities and dimensionality reduction through the MMS algorithm.

Original Source(s)

Multimodal breast cancer diagnosis using feature fusion and deep learning

  • by Varun Malik, Tahani Alsubait, Mudassir Khan, Upasana Lakhina, Alaa Menshawi, Stuti Mehla, Loveleena Mukhija, Meteb Altaf, Leila Jamel

  • July 21, 2026

Related Content

Molecular characteristics and advances in precision therapeutics for HR-positive and HER2-low breast cancer

Parasite in cancer therapy: molecular mechanisms and translational potential

Preliminary exploration on using entropy-weighted hybrid pooling in CNN for ultrasound breast cancer detection