Please use this identifier to cite or link to this item: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4540
Title: User Emotion based Deepmoji Desktop Partner
Authors: SALLAY, N.S.
Issue Date: 12-Aug-2021
Abstract: The user emotion based deepmoji desktop partner is developed to address the emotional awareness of virtual assistants. Developing a response generation model which gives responses according to the emotions of the text is the solution proposed. Despite of being a computer application, people do not accept “google assistant” or “siri” as a friend because they lack in emotional intelligence and empathetic interactions in conversations. Emotional intelligence of computers has gained a notable attention of researchers because of its extensive application esteem and high potential in developing a tremendous market value. Sentiments are intrinsic value of human beings and humans have the ability of identifying others sentiments adopt the conversations, according to the emotions of the interacting parties. The proposed system develop a model which identifies emotion of the user by using the input text and generate response according to the emotion captured. People identify emotions by facial expressions, gestures and intensity of voice tone etc. The proposed solution is focused to develop OS independent AI desktop partner to interact with users based on their emotions, captured through facial expressions and texts. According to the review on past attempts on achieving emotional intelligence, technologies and different processes on achieving emotional awareness , Facial Expression Recognition and Sentiment Analysis techniques were used for the development and Deepmoji is used to train the AI model to classify the sentiments behind the phases. Accuracy of the emotions directly effect the empathetic characteristic of the conversation. Therefore capturing the correct emotion is a crucial task. The system implementation has two main modules, emoji module which is the user interface of the desktop partner and the chat module. Group of IT professionals and non-IT professionals were selected to evaluate the application and results of a validated questionnaire containing questions based on performance, emoji accuracy, appropriateness with the comparison of other existing virtual assistants were evaluated. Questionnaire was evaluated using Cronbach’s Alpha test resulting Cronbach’s Alpha at 0.75 which is a higher positive level. The result of the evaluation indicates that the proposed concept is a proper software solution for the problem of emotionally aware virtual assistants.This dissertation includes the methodology, design, implementation and the evaluation of the implemented system prototype.
URI: http://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4540
Appears in Collections:2020

Files in This Item:
File Description SizeFormat 
2017 MCS 071.pdf10.64 MBAdobe PDFView/Open


Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.