A New Efficient Classification Method for Monkeypox Virus Skin Lesions Based on MPFormer

Monkeypox is a serious zoonotic disease caused by the monkeypox virus. Traditional laboratory detection methods are time-consuming and have high biosafety requirements; in particular, electron microscopy is expensive and complicated to operate. To address these challenges, this paper proposes a novel classification algorithm for monkeypox virus skin lesions based on the Vision Transformer-MPFormer. We adopted a transfer learning strategy to train the MPFormer model on a monkeypox virus skin lesion dataset for accurate classification. Experimental results on the MSLD dataset demonstrate that the proposed algorithm achieves significant performance improvements compared to common models and other classification algorithms, with accuracy, precision, recall, and F1-score reaching 97.78%, 97.88%, 97.78%, and 97.78%, respectively. This study provides an efficient computer-aided diagnostic tool for rapid and accurate detection of the monkeypox virus, offering important practical application value.