Apparatus, method and program for selecting images for a slideshow from among a large number of captured images

A method of processing information in an information processing apparatus selecting a predetermined number of images from a plurality of images captured at an event, the method comprising the steps of:

• by the information processing apparatus,
• analyzing the plurality of images, and calculating individual evaluation values of the images;
• on the basis of a time-difference value indicating a difference between shooting dates and times of adjacent images when the plurality of images are rearranged in order of shooting date and time, first clustering to determine a first boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• on the basis of a distance-difference value indicating a difference between shooting positions of adjacent images when the plurality of images are rearranged in order of shooting date and time, second clustering to determine a second boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• sorting the plurality of images rearranged in order of shooting date and time into the plurality of sub-events in accordance with the determined first and second boundary groups;
• image-selecting to determine images to be selected which includes excluding images having small evaluation values among the images sorted for each of the sub-events by the sorting step,
• in which the image-selecting step is to determine the images to be selected in accordance with (i) precedence of person images, (ii) precedence of scenic images, and (iii) normal, as a principle of selecting images;
• and
• specifying to allow a user to specify a desired one of (i) the precedence of person images, (ii) the precedence of scenic images, or (iii) the normal, as the principle of selecting images by the image-selecting step, in which when the precedence of person images principle is specified the image-selecting step will give priority to the person images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, when the precedence of scenic images principle is specified the image-selecting step will give priority to the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, and when the normal principle is specified the image-selecting step will give equal priority to the person images and the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, and
• in which each evaluation value is obtained by combining a plurality of calculated evaluation values, in which a first calculated evaluation value is calculated in accordance with sizes, positions and expressions of faces in a respective image and number of people in the respective image, and a second calculated evaluation value is calculated in accordance with composition of scene in the respective image.

A non-transitory computer-readable medium having stored thereon a program for controlling an information processing apparatus selecting a predetermined number of images from a plurality of images captured at an event, the program causing a computer of the information processing apparatus to perform processing comprising the steps of:

• analyzing the plurality of images, and calculating individual evaluation values of the images;
• on the basis of a time-difference value indicating a difference between shooting dates and times of adjacent images when the plurality of images are rearranged in order of shooting date and time, first clustering to determine a first boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• on the basis of a distance-difference value indicating a difference between shooting positions of adjacent images when the plurality of images are rearranged in order of shooting date and time, second clustering to determine a second boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• sorting the plurality of images rearranged in order of shooting date and time into the plurality of sub-events in accordance with the determined first and second boundary groups;
• and
• image-selecting to determine images to be selected which includes excluding images having small evaluation values among the images sorted for each of the sub-events by the sorting step,
• in which the image-selecting step is to determine the images to be selected in accordance with (i) precedence of person images, (ii) precedence of scenic images, and (iii) normal, as a principle of selecting images;
• and
• specifying to allow a user to specify a desired one of (i) the precedence of person images, (ii) the precedence of scenic images, or (iii) the normal, as a principle of selecting images by the image-selecting step, in which when the precedence of person images principle is specified the image-selecting step will give priority to the person images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, when the precedence of scenic images principle is specified the image-selecting step will give priority to the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, and when the normal principle is specified the image-selecting step will give equal priority to the person images and the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting step, and
• in which each evaluation value is obtained by combining a plurality of calculated evaluation values, in which a first calculated evaluation value is calculated in accordance with sizes, positions and expressions of faces in a respective image and number of people in the respective image, and a second calculated evaluation value is calculated in accordance with composition of scene in the respective image.

An information processing apparatus selecting a predetermined number of images from a plurality of images captured at an event, the information processing apparatus comprising:

• an image analyzing mechanism analyzing the plurality of images, and calculating individual evaluation values of the images;
• on the basis of a time-difference value indicating a difference between shooting dates and times of adjacent images when the plurality of images are rearranged in order of shooting date and time, a first clustering mechanism determining a first boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• on the basis of a distance-difference value indicating a difference between shooting dates and times of adjacent images when the plurality of images are rearranged in order of shooting date and time, a second clustering mechanism for determining a second boundary group including a plurality of boundaries for sorting the plurality of images rearranged in order of shooting date and time into a plurality of sub-events;
• a sorting mechanism sorting the plurality of images rearranged in order of shooting date and time into the plurality of sub-events in accordance with the determined first and second boundary groups;
• and
• an image selecting mechanism determining images to be selected which includes excluding images having small evaluation values among the images sorted for each of the sub-events by the sorting mechanism,
• in which the image selecting mechanism is to determine the images to be selected in accordance with (i) precedence of person images, (ii) precedence of scenic images, and (iii) normal, as a principle of selecting images;
• and
• a specifying mechanism allowing a user to specify a desired one of (i) the precedence of person images, (ii) the precedence of scenic images, or (iii) the normal, as the principle of selecting images by the image selecting mechanism, in which when the precedence of person images principle is specified the image selecting mechanism will give priority to the person images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting mechanism, when the precedence of scenic images principle is specified the image selecting mechanism will give priority to the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting mechanism, and when the normal principle is specified the image selecting mechanism will give equal priority to the person images and the scenic images for selection when excluding images having small evaluation values from among the images sorted for each of the sub-events by the sorting mechanism, and
• in which each evaluation value is obtained by combining a plurality of calculated evaluation values, in which a first calculated evaluation value is calculated in accordance with sizes, positions and expressions of faces in a respective image and number of people in the respective image, and a second calculated evaluation value is calculated in accordance with composition of scene in the respective image.