Revolutionizing employee performance evaluation: An automated approach with machine vision and ensemble learning for precision monitoring and efficient decision-making

Employee evaluation has been used in all industries as a strategic approach to align company aims and objectives efficiently and effectively. Individual employee performance in the task assigned to them needs to be monitored to keep track of progress and for decision making by managers. Traditional method of monitoring is prone to bias, nepotism, and high error with high time consumption. This project proposes a solution to automate the monitoring process by using machine vision and an ensemble approach machine learning models with an object detection model. Through this approach an action recognition pipeline is built that has an overall accuracy of 94.23% and precision of 91.94% while being able to process a minimum of 15 frames per second using a normal webcam. The monitored data is then visualized in a web dashboard for the managers to assess at any time. Using this system, better insight is possible into the employee’s work habits and monitoring is made easier and more efficient for the managers and supervisors.