Based on the result of the above determination, the person extraction unit 10 calculates an appearance frequency for each person (for each person ID) detected in the moving image data to be analyzed. As the appearance frequency, for example, the number of appearances (the number of pieces of child data in which a person appears) or an appearance rate (=(the number of pieces of child data in which a person appears)/(the number of pieces of child data in the moving image data 100 to be analyzed)) can be considered.

After calculating the appearance frequency, the person extraction unit 10 extracts a person (person ID) satisfying a predetermined condition based on the calculated appearance frequency. As the predetermined condition, any condition expressing a high appearance frequency may be adopted. For example, the predetermined condition may be “an appearance frequency of equal to or higher than a predetermined level”. Specifically, “the number of appearances of equal to or higher than a predetermined number of times” and “the appearance rate of equal to or higher than a predetermined rate” may be used.

In addition, the predetermined condition may be “a predetermined number of persons in descending order of appearance frequency”. Specifically, the predetermined condition may be “within a predetermined rank in the ranking in descending order of appearance frequency”. The predetermined conditions exemplified herein are examples and the present invention is not limited thereto.