Effect of Image Resolution on Remote Photoplethysmography: Towards Emotion Detection in Children with Autism Spectrum Disorder

Abstract

Aiming to stimulate social abilities in children with Autism Spectrum Disorder was developed at UFES/Brazil a robot to interact with them. One of the functions that are being developed in that robot is emotion recognition using cardiac activity. Traditional heart rate measuring methods normally use contact sensors, often uncomfortable for autistic children. So, a method to measure it remotely and online is an important contribution to this field of research. This paper presents a method to remotely measure heart rate using a low-cost RGB camera. The effect of the different image resolutions in the final result is evaluated, aiming in a future online system, which needs a small computation time. The results indicate that the resolution does not have a relevant influence on measure's precision, so, the resolution of the video can be reduced without a big effect on the error.