|Title:||Exploring Affective Computing for Enhancing the Museum Experience with Online Collections|
|Authors:||Gunho Chae, S. Joon Park, Robert Stein, Jungwha Kim, Susan Wiedenbeck|
|Publication:||MW2012: Museums and the Web 2012|
Traditionally, the focus of online museum collections has been to facilitate access and discovery for subject category experts seeking information about objects in the collection (Rayward & Twidale, 1999). Today, museums observe an increasing number of 'untrained eyes' visiting their online collections to browse or search art collections for intrinsic enjoyment and leisure (Marty, Sayre, & Fantoni, 2011; Packer, 2010). Research by the Steve.museum social tagging project has demonstrated that the language used by novice visitors to describe works of art differs significantly from that descriptive metadata used by museums to describe the same collections (Trant, 2009). Many visitors to museums' online collections seek experience-centered online experiences that can be fun and pleasurable, rather than a suite of advanced features.
However, most of today's online collections do not fully support such an experience-centered environment. A quick survey could show that these systems lack not only functionality, but also the appropriate affective and hedonic aspects necessary to support such an experience (Blythe, Wright, McCarthy, & Bertelsen, 2006). It is necessary to take holistic approaches for online museum experiences, which go beyond accessibility and functionality, to address the aspects of emotional interaction (Johnson, Gardner, & Wiles, 2004).
Using the online collection of the Indianapolis Museum of Art (IMA), this study seeks to design and implement an affective online search and retrieval system for visitors who may have general interests in art, but may not know where to start. Techniques from the affective computing literature will be applied to develop an affective embedded agent that can recognize and respond to the users' expressed emotions and preferences. Social tagging tools will be used to record the visitors' subjective views and interests in arts. These social tags will inform the affective agent and allow it to interact with visitors based on this information.
Because the proposed agent uses interactive techniques to incorporate the users' emotions and preferences during the image retrieval process, the untrained online visitors can easily refine their interests about art collections and map those interests to terminology and metadata maintained by the museum. As such, online visitors do not need to be subject matter experts with knowledge in art historical terminology to navigate art collections; an enjoyable browsing experience can be created to support visitors' known and unknown preferences. This study can contribute to the field of museum technology by providing empirical guidelines on how an affective agent can be designed and influence the internal state of a person to improve overall museum experience in an online setting.
When combined with user-generated social tags, an embedded agent can be a powerful search method that provides 'general-purpose question-answering capabilities' in the context of online museums (Etzioni, 2011). We hope our research serves to broaden our understanding of the role that emotions play in use of search and retrieval systems by novices, and that it will provide useful guidelines for the online museum community in studying interactions with systems that are intelligent and sensitive to online visitors' emotions and preferences.