Affective TV: Concepts of Affective Computing Applied to Digital Television
Abstract: The traditional broadcast TV viewing experience has barely evolved since its inception, remaining mostly static despite many technical advances. Smart TVs show attempts of filling this gap, but present challenges, such as limiting functionalities to specific models and lack of standardization. Privacy concerns arise as smart TVs connect to advertising and monitoring services. In the spectrum of interactivity, an option that stands out is affective computing, an interdisciplinary field that seeks to develop systems capable of recognizing, expressing and responding to human emotions. This work proposes the incorporation of affective computing techniques and concepts to improve the experience and interactivity with digital TV, naming it “Affective TV”. The work presents a modular architecture, recognition modules developed for multiple modes of interaction and a fully operational implementation of the architecture, developed for the standard digital TV middleware in Brazil, Ginga. Affective TV uses audio and video capturing devices and allows users to set up their environments. Recognition modules capture and classify data, communicating directly to the TV middleware. Proof-of-concept applications, incorporating voice and hand pose interactions with facial expression recognition, were evaluated using the GQM. UEQ-S and TAM questionnaires were employed. Very positive results were obtained, including an excellent UEQ rating, showcasing technical feasibility, attractiveness, user experience, perceived usefulness, and ease of use. The proposal enriches the digital TV experience, providing a novel, interactive model with user-centric customization and emotion-driven responses.
Loading