Pixel-guided pattern alignment based Hopfield Neural Networks for generalize cancer diagnosis
Biomedical Signal Processing and Control, cilt.103, 2025 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 103
- Basım Tarihi: 2025
- Doi Numarası: 10.1016/j.bspc.2024.107397
- Dergi Adı: Biomedical Signal Processing and Control
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, EMBASE, INSPEC
- Anahtar Kelimeler: Cancer diagnostics, Contour predictor, Global pixels, Local pixel, Profile predictor, Texture improvement
- Ankara Hacı Bayram Veli Üniversitesi Adresli: Evet
Özet
Early diagnosis of cancerous tissues in radiology images is challenging due to subtle changes in fine details. This study introduces a novel approach using Hopfield Neural Networks (HNN) for texture enhancement and cancer detection. The method employs recursive filtering to separate noise from essential details, preserving geometric features. An innovative energy function in HNN, integrating optimal global (β G) and local (β L) pixels, facilitates accurate prediction of cancerous patterns. Comprehensive experiments demonstrate the method's superior performance, achieving an accuracy of 99.92%. The results highlight the potential of this enhanced HNN-based approach in providing reliable early cancer diagnosis by effectively separating noise and accurately predicting cancerous patterns. Integration of Ant Colony Optimization (ACO) further enhances detection capabilities, leading to a higher probability of correctly identifying cancerous cells, offering a promising tool for radiologists in early cancer detection and treatment. The proposed method effectively distinguishes cancerous tissues from non-cancerous ones across diverse datasets, showing its robustness and potential for clinical applications. Future recommendations include exploring the application of the proposed method with other neural networks, such as recurrent neural networks (RNNs).