Investigated users’ needs by keeping track of use cases and interviewing from 5 different churches to increase efficiency in development phase

Implemented user authentication with JWT, RESTful API and Django REST framework

Developed database with PostgreSQL to save storage space for event and user information by 5%

Applied Redux as state management tool to reduce time needed in system maintenance

Deployed the system on cloud using Heroku to avoid the higher cost of local server

Recognized user’s facial expression, hand gesture and verbal information by Intel RealSense, and designed algorithms convert them to emotional values for multimodal information analysis

Implemented non-verbal affective computing with PAD emotional model, and increased the accuracy by 18% by calculating different weights of emotional values

Implemented fusion of PAD emotional values on decision layer using Linear Regression method and designed algorithm for mapping them in three-dimensional emotional space