Gene-related Parkinson's disease diagnosis via feature-based multi-branch octave convolution network

Haijun Lei, Yuchen Zhang, Hancong Li, Zhongwei Huang, Chien-Hung Liu, Feng Zhou, Ee-Leng Tan, Xiaohua Xiao, Yi Lei, Huoyou Hu, Yaohui Huang, Baiying Lei

Abstract: Highlights•A joint learning framework called feature-based multi-branch octave convolution network (FMOCNN) for gene-related cohort PD diagnosis.•The cardinality-constrained sample feature selection (CCSFS) model selects informative samples and features.•A multi-branch octave convolution neural network (MBOCNN) is established to reduce redundant feature and enhance the feature expression ability.