A rotation meanout network with invariance for dermoscopy image classification and retrieval

Yilan Zhang, Fengying Xie, Xuedong Song, Hangning Zhou, Yiguang Yang, Haopeng Zhang, Jie Liu

Abstract: Highlights•A rotation meanout (RM) network is proposed to extract rotation-invariant features.•RM can be flexibly embedded in convolutional neural networks (CNNs).•The proposed method does not change the network structure or increase the parameters.•RM improves the performance of CNNs in skin disease classification and retrieval.