Mpox-PyramidTransferNet: A Hierarchical Transfer Learning Framework for Monkeypox and Dermatological Disease Classification

Juanxi Tian, Kaixin Zhu, Wentao Fan

Abstract: During the COVID-19 pandemic, the global immunity of human populations was adversely affected, posing threats for other contagious diseases. Monkeypox cases have surged recently, bringing new challenges to global public health. In this study, we propose Mpox-PyramidTransferNet, a novel hierarchical transfer learning architecture equipped with various attention mechanisms for accurate classification of monkeypox and other skin conditions. Our model takes advantage of the pyramid network composed of InceptionResnet-V2, ResNet152V2 and DenseNet121, which have been pre-trained on large-scale image recognition tasks, thereby extracting multi-scale features. The classification performance is further improved by the attention mechanisms. Furthermore, we fine-tune the model with monkeypox datasets. The performance of our proposed method is evaluated on the latest MSLDv2 dataset and compared with several state-of-the-art methods. Results show validation and classification accuracy of Mpox-PyramidTransferNet exceeds 98.5%, significantly outperforming other models. The code of our model is available at: https://github.com/tianshijing/Mpox-PyramidTransferNet.