Hybrid Efficient-KAN: A Multimodal Deep Learning Framework for Choroidal Tumor Classification Hybrid Efficient-KAN: Koroid Tümörlerinin Siniflandirilmasi için Çok Modlu Bir Derin Ögrenme Çerçevesi
34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/siu71813.2026.11636640
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: choroidal neoplasms, EfficientNet, kolmogorov-arnold networks (KAN), tumor classification
- Atatürk Üniversitesi Adresli: Evet
Özet
In this study, a novel multimodal deep learning approach is proposed to address the challenges of missing modalities and class imbalance in the classification of rare choroidal tumors. In the proposed Hybrid Efficient-KAN model, the parameter-efficient EfficientNet-B0 architecture is employed for feature extraction, while Kolmogorov-Arnold Networks (KAN) with learnable activation functions are utilized in the classification stage. A multimodal dataset consisting of angiography and ultrasonography images is used, and the model performance is evaluated through comprehensive ablation studies under various data processing strategies. It is demonstrated by the experimental results that an accuracy of 94.73% is achieved and superior performance is obtained compared to conventional CNN-based methods. It is further shown that diagnostic accuracy is enhanced through multimodal KAN-based classification, and an effective solution to the problem of missing data is provided.