Deep Learning Approaches for Sentiment Analysis in Instagram Images

Authors

Abstract

Sentiment analysis in images is the discovery of the emotion that an image passes to a person when looking at it. Nowadays social media is used by millions of people and most of the available platforms use images as one of their main communication features. In this work, we use Convolutional Neural Networks (CNN) to determine the sentiment and to improve the user experience when posting an image to social media, specifically to Instagram. Different CNN are proposed and evaluated to perform this task. A mobile app was developed to interact with the user and provide a simple service to predict the sentiment of an image before the user posts it to Instagram.

Conference

RECPAD 2019 - Portuguese Conference on Pattern Recognition, October 2019