Recognition of Advertisement Emotions with Application to Computational Advertising

Abstract

Advertisements (ads) often contain strong affective content to capture viewer attention and convey an effective message to the audience. However, most computational affect recognition (AR) approaches examine ads via the text modality, and only limited work has been devoted to decoding ad emotions from audiovisual or user cues. This work (1) compiles an affective ad dataset capable of evoking coherent emotions across users; (2) explores the efficacy of content-centric convolutional neural network (CNN) features for AR vis-\~a-vis handcrafted audio-visual descriptors; (3) examines user-centric ad AR from Electroencephalogram (EEG) responses acquired during ad-viewing, and (4) demonstrates how better affect predictions facilitate effective computational advertising as determined by a study involving 18 users. Experiments reveal that (a) CNN features outperform audiovisual descriptors for content-centric AR; (b) EEG features are able to encode ad-induced emotions better than content-based features; (c) Multi-task learning performs best among a slew of classification algorithms to achieve optimal AR, and (d) Pursuant to (b), EEG features also enable optimized ad insertion onto streamed video, as compared to content-based or manual insertion techniques in terms of ad memorability and overall user experience.