I. Introduction
Reading the human mind is one of the most attracting problems for brain-computer interface (BCI) research. Over the past decades, researchers have actively been studying how to decode the human mind from brain signals to facilitate communication or restore past memories [1]–[5]. For example, BCI systems are developed to help people with disabilities to communicate with others by recognizing what they are thinking [6]. In addition, researchers have paid much attention to interpreting visual information in the human mind and identifying the visual stimulus that a human has experienced from brain signals [7]. Recent studies have shown promising performance in recognizing the category of visual stimulus based on the brain signals [8], [9].