Video based attention assessment

Abstract

Attention Assessment is one of the most interesting topic of research in the ongoing era. Many researches going on in this eld to develop a tool that can accurately measure the attention level. The past researches on this topic had shown a strong relationship between heart rate variability (HRV), heart rate (HR), and cognitive state of mind (Stress/Attention). Researchers have used traditional heart rate calculation methods like using ECG signal and photoplethysmography (PPG) signal in the previous approaches. However, the traditional approaches of nding heart rate are not user-friendly, as the calculation of ECG signal involves a complex procedure; at the same time, these methods are not cost-friendly. In recent time, many researchers are trying to measure heart rate using video-based remote PPG extraction methods; these methods do not require any contact and can be measured remotely so that they can be bene cial in elds like telemedicine. In the past, researcher has tried di erent methods to nd heart rate using the video-based approach, some have used a deep learning-based approach, and others have used signal processing-based approaches. At the same time, there is no signi cant research on nding the attention level from scratch using a video-based approach is there till now. This research aims to excess the strengths and weaknesses of these existing models and theories and tries to build a video-based attention assessment system using computer vision and signal processing, which will be reliable and at the same time cost-e ective. We also intend to make a platform(Android App/Web extension) that will increase the accessibility of our product in day-to-day life.