Measuring Cognitive Load in Children's Videos Using Motion Estimation

Young children are particularly sensible to cognitive load, as the quantity and complexity of presented content significantly impact their learning capacity. To ensure a better learning experience, educational materials must be carefully designed to minimize cognitive overload. This study focuses on videos available on online streaming services that are targeted at children. A significant part of these videos is not created by educational experts, and a simple search can lead children to videos that may be cognitively overwhelming. This study aims to introduce a novel approach to assess the cognitive load of videos and, thus, determine their suitability for children by employing a motion detection approach. More precisely, we have implemented a block-matching motion detection approach to evaluate the overall motion present in a video. This approach was utilized to process a dataset of videos, facilitating the training of a classifier and the determination of a threshold that clearly separates high-motion videos from low-motion videos. The classifier has been used to develop a user-friendly tool designed to assist parents in evaluating videos before allowing their children to watch them.