For example, by setting the time width of the time window to a desired state, it is possible to obtain the extraction result of a person frequently appearing at desired time intervals. For example, by setting the time width of the time window to one hour, it is possible to obtain the extraction result of a person having a high appearance frequency for each one-hour period. In addition, by setting the time width of the time window to 1 day, it is possible to obtain the extraction result of a person having a high appearance frequency every day.

In addition, by individually setting the start position and the end position of each of the plurality of time windows, it is possible to obtain the extraction result of a person frequently appearing at a desired timing. For example, by setting a plurality of time windows so as to include each of a plurality of timings at which a crime, such as pickpocketing, molesting, and surreptitious shooting, has occurred, it is possible to obtain the extraction result of a person frequently appearing at the timing of the crime.

Here, an application example of the present example embodiment will be described. For example, in a case where similar crimes (pickpocketing, molesting, surreptitious shooting, shoplifting, or the like) have frequently occurred at a certain place, moving image data captured by a security camera installed at the place is provided to the data processing apparatus 1 as data to be processed.

Then, the operator sets the start position and the end position of each of the plurality of time windows so as to include the occurrence time of each of the plurality of crimes. As a result, a person having a high appearance frequency at a plurality of crime occurrence times (a person who is frequently present at the crime site during a plurality of crime occurrence times) is extracted.