Design a proposal of an Affective Computing technology.
Emotions: they can say a lot about a person. The potential of emotions in the realm of technology and Human-Computer Interaction is enormous, giving a further possibility of identifying a relationship between man and machine. The field of Emotion Recognition is offering different paths to discover with extremely heterogeneous, interesting and fruitful outcomes.
Credit also goes to: Sara Fazzini.
In a world where communication is spreading, marketing agencies are dealing with the effort of designing the most e(a)ffective advertisements. In this process, the role of emotions felt by the user in response to viewing a certain TV commercial is crucial. Project AMSTOS is an Affective Multimedia retrieval, content Saliency mapping, Transdermal Optical imaging Software which uses Emotion Recognition to correctly predict emotion elicitation and effectiveness of TV commercials.
The software is based on a Convolutional Neural Network and Deep Learning to complete a process of feature extraction, which is imported in a 2-dimensional Arousal-Valence Space.
After having generated the analysed video's Affect Curve, through a specific target group's Content Saliency Mapping and Transdermal Optical Imaging values, it is possible to determine the effectiveness of the TV commercial by comparing the machine-recognized emotions with the emotions elicited in the users and the those that the marketing agency wanted to convey through the advertisement. To read the project proposal please contact me directly.
