Digital imaging and multimedia technology from the optoelectronic and information science fields are combined to invite people to engage in the scientific model in a relaxed and interesting way. The purpose of this study is to embed multiple color images into one color image using neural network technology and color harmony. The neural network system uses a hierarchical pattern detection scheme or system, which includes both low- and top-level pattern detection systems. The low-level system is responsible for and capable of detecting sub-patterns, while the top-level system detects patterns and outlines and identifies the classes to which the patterns belong. Binary images were obtained from the originals, and from the binary images pixelated images were obtained. We developed a new method to process experimental data and evaluated the effect of signal to noise ratio in PC-based digital signal processing system. This novel device can be operated automatically and economically in a participle display of the scientific model. Furthermore, concealed information can be inserted, and sound and other effects used to attract the attention of visitors.