Automatic Emoji Insertion Based on Environment Context Signals for the Demonstration of Pervasive Computing Features

Abstract

Pervasive computing applications are becoming much closer to take over the computing environment in different areas through different portable and smart devices. This includes healthcare, education, military, retail, instant messaging, and much more. For the vision to be a reality, the use of context-awareness as a technique for the development of pervasive computing applications is of growing interest. Context-awareness gives flexibility, adaptability, and autonomy to the computing system. In this paper, design and development of a MATLAB-based instant messaging tool with automatic emoji insertion is presented. This tool incorporates a two-way application that uses ThingSpeak as a middleware. It also takes inputs from different sensors that indicate a specific context on which a decision is made about the type of emoji to be inserted into the text message. The contexts that are taken into consideration are physical, human, and local and networked ICT environments. In addition, the system demonstrates the use of artificial intelligence for emotion detection as well as physical environment tagging. After implementation, we analyzed its accuracy, response time, and major constraints that need to be taken into consideration. Most of the contexts were detected with high accuracy between 86.6667% and 100%, in addition to faster response times between 565.28ms and 1.743s when tested individually. However, in a scenario where all contexts are being used in a prioritized way, the performance of the overall system was affected by the 66.67% medium accuracy of the emotion detection system(intelligent feature) and its slow response times of 5.979s. Further improvements on the context detection systems and their constraints will highly improve the overall system accuracy and response times. This work has an educational aspect of demonstrating multiple types of context awareness in a practical pervasive computing application.