System, program, and method for generating visual-guidance information

摘要

Exemplary embodiments of the present invention provide a system to generate visual-guidance information in which the system can be reduced in size and cost and which ensures a proper eye flow. A layout unit operates in such a way that a layout image is segmented into multiple regions; the saliency features and positional information of the partial regions are calculated from image data, and the calculated saliency features are weighted. A starting-point region is then detected from among the multiple partial regions from the calculated saliency features and the reference points of the partial regions are detected from a visual-guidance feature space based on the weighted saliency features and the calculated positional information. Then the distance between the detected reference points is calculated and visual-guidance information is generated from the calculated inter-reference-point distance. The visual-guidance information is expressed as a visual-guidance path that passes through the partial regions in sequence, with a starting-point region as a starting point and a partial region having a shortest inter-reference-point distance from the preceding partial region as the following partial region.