Fig. 1From: Video person reidentification based on neural ordinary differential equations and graph convolution networkFeature extraction of continuous hidden state modelBack to article page