Detecting emotion in images using deep learning for sentiment analysis

Ting, Wai Lok (2017) Detecting emotion in images using deep learning for sentiment analysis. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full text) Detecting emotion in images using deep learning for sentiment analysis.pdf - Submitted Version Download (1MB)

Abstract

Anger, contempt, disgust, fear, joy, sadness, and surprise are the emotions that can be feel by humans. These seven basic emotion are the results of internal chemical process in the brain and also from the influence of external environment factors. Due these emotions happen in the inner process of the brain, it is very hard to quantify it and have a definitive definition for emotion. In everyday life, emotions impact human’s judgments and decision in different aspect. These mental states experienced by human can affect human along with decision making, facial expression and thoughts. Thus, a computer that are able to recognize facial expression will be very useful providing analytics and insight for better decision making. The major objective of this study is to develop a model that is able to identify 7 emotion from images using various AI techniques like deep learning and image processing. Seven groups of images, each consisting 7 basic emotion human facial expression respectively were selected to be used in the training of the model. The scope of this study only focus on identifying facial expression of human. Iterative development model was used throughout the development of this project. The performance test for the classifier model reaches 77% accuracy. This study is expected to have vast potential and contribution in real life assisting different domain like analytics, sentiment analysis, improved decision making, etc.

Actions (login required)

View Item