Optimizing vision transformer for melanoma classification: integration of feature selection and InfoNCE loss
Information Sciences|更新时间:2025-09-28
|
Optimizing vision transformer for melanoma classification: integration of feature selection and InfoNCE loss
“In the field of melanoma image classification, researchers have proposed an improved Vision Transformer model that significantly improves classification accuracy and diagnostic efficiency through dynamic feature screening and contrastive learning.”
Optics and Precision EngineeringVol. 33, Issue 16, Pages: 2649-2660(2025)
HUANG Jinjie,MA Yuanxue.Optimizing vision transformer for melanoma classification: integration of feature selection and InfoNCE loss[J].Optics and Precision Engineering,2025,33(16):2649-2660.