Social and Content Hybrid Image Recommender System for Mobile Social Networks

Abstract

One of the advantages of social networks is the possibility to socialize and personalize the content created or shared by the users. In mobile social networks, where the devices have limited capabilities in terms of screen size and computing power, Multimedia Recommender Systems help to present the most relevant content to the users, depending on their tastes, relationships and profile. Previous recommender systems are not able to cope with the uncertainty of automated tagging and are knowledge domain dependant. In addition, the instantiation of a recommender in this domain should cope with problems arising from the collaborative filtering inherent nature (cold start, banana problem, large number of users to run, etc.). The solution presented in this paper addresses the abovementioned problems by proposing a hybrid image recommender system, which combines collaborative filtering (social techniques) with content-based techniques, leaving the user the liberty to give these processes a personal weight. It takes into account aesthetics and the formal characteristics of the images to overcome the problems of current techniques, improving the performance of existing systems to create a mobile social networks recommender with a high degree of adaptation to any kind of user.

On the influence of low-level visual features in film classification

“…Besides, these studies have been completed with other approaches regarding the sensory experience evoked by images (Gombrich [22]) as well as human influence and computation of color (Itten [23] and Davis [24]). We already took into account these approaches when working with pictures for still image characterization, which we developed in previous research works [25].…”

“…Our model's descriptors in this first block are based on previous works of the authors [25][31], where their effectiveness has been proved. These descriptors, as it has been underlined, are based in Aumont's film studies (focused on aesthetic studies), although some of them are also similar to the technical descriptors defined in MPEG-7 [32][26].…”

MPEG-7 low level image descriptors for modeling users' web pages visual appeal opinion

“…Our model is strictly based on the analysis of the low level image characteristics of the website screenshots [18]. To do so, we make use of a set of parameters derived from the MPEG-7 standard, which are related to three main dimensions of the image: luminance, texture and color, as some of the most important properties for the visual psychobiological process ( [19], [20]):…”

Estudo da percepção humana e mineração de preferências contextuais na recomendação de imagens para o usuário

“…Segundo (HERLOCKER et al, 2004), são três as principais etapas que compreende um sistema de recomendação, sendo elas: os dados de entrada dos clientes adquiridos através da preferência dos clientes, cálculo da recomendação por meio de técnicas apropriadas e a apresentação dos resultados obtidos com a recomendação para os clientes. Sistemas de recomendações são utilizados para fazer diversos tipos de recomendação como, por exemplo, de livros (LINDEN;SMITH;YORK, 2003), vídeos (DAVIDSON et al, 2010, páginas da web (BALABANOVIĆ; SHOHAM, 1997), imagens (SANCHEZ et al, 2012), músicas (KUO et al, 2005) e notícias (LIU; DOLAN; PEDERSEN, 2010).…”

“…O algoritmo apresentando depende dos dados cadastrais dos usuários que precisam estar todos preenchidos. Sanchez et al (2012) propõem um sistema de recomendação híbrido que considera a estética e as características formais das imagens para fazer recomendações de imagens em redes sociais móveis, com alto grau de adaptação a qualquer tipo de usuário e interface. O sistema é composto por dois processos.…”