An end-to-end deep generative approach with meta-learning optimization for zero-shot object classification

Xiaofeng Xu, Xianglin Bao, Xingyu Lu, Ruiheng Zhang, Xinquan Chen, Guifu Lu

Abstract: Highlights•An end-to-end deep approach is proposed for zero-shot object classification.•Proposed approach trains the data generator and object classifier jointly.•A meta-learning optimization is designed for the projection domain shift problem.•Experimental results demonstrate the significant superiority of proposed approach.