How do users make a people-centric slideshow?

Abstract

This paper presents a pilot user study that attempts to shed light on the ways users create people-centric slideshows, with the objective of scaling it up to a crowdsourcing experiment. The study focuses on two major directions, namely image selection and image sequencing. Participants were asked to select photos of a specific person from an initial set and arrange them into a slideshow. Results show that there is correlation between specific predictors and selected images, as well as their relative position in the final sequence. This indicates that a crowdsourcing experiment will indeed highlight the characteristics of the average user, which can then be incorporated into an automatic people-centric slideshow creator.

Modeling Image Appeal Based on Crowd Preferences for Automated Person-Centric Collage Creation

“…Analysis of the crowdsourcing results showed that the average and median number of selected images over all albums were 17.05 and 13 respectively, which is in accordance with the previous findings of [18]. Another interesting result is that all 60 images of all 5 albums were selected at least once, i.e., there were no images that were never selected.…”

“…Examination of how IA influences selections in photo sequences and correlates with users' visual attention patterns is presented in [6]-however, no attempt is made to automatically replicate observed image selections in this work. A pilot-study employing 14 users to identify low and high-level image attributes influence image selections on a child -themed photo album is described in [18].…”

“…All the aforementioned studies, with the exception of [18] (which is limited in terms of number of users and album themes), focus on general photo-collections, that may include people, scenery, pets, food, objects etc. However, it has been shown that the majority of personal photocollections comprise images portraying people and faces, and such images have been found to be most popular in social media [2].…”