Premium
The effect of demoting near‐duplicate pictures
Author(s) -
Bergman Ofer,
Tucker Simon,
Dahamshy Somaya
Publication year - 2018
Publication title -
proceedings of the association for information science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.193
H-Index - 14
ISSN - 2373-9231
DOI - 10.1002/pra2.2018.14505501103
Subject(s) - computer science , stack (abstract data type) , presentation (obstetrics) , information retrieval , artificial intelligence , computer graphics (images) , medicine , radiology , programming language
People often take several pictures of the same situation “just in case something goes wrong”. As a result, their picture folders tend to be cluttered with near‐duplicate pictures. These near‐duplicate pictures reduce the efficiency of retrieving a specific picture as they compete for the users' attention. To address the problem of near‐duplicates, we developed Duplipix , a working prototype that identifies near‐duplicate pictures and piles them in a stack. When a user hovers the mouse over a stack, it opens allowing the user to select one or more of the near‐duplicate pictures. In an initial within‐subjects evaluation, we asked our 82 participants to retrieve a random picture from the same collection ‐ with and without DupliPix . The use of DupliPix resulted in significantly faster picture retrieval, and the majority of participants indicated that using it is easier, more aesthetic and interesting than the standard picture presentation. They would also like DupliPix to be included in their next operating system.