The human eye can only perceive signals within a specific frequency range. However, studying signals beyond this range can provide valuable insights, especially when studying human behaviour in response to a stressful situation. To visualise these spatio-temporal variations, video magnification techniques have been developed. When blood flows through the skin, it undergoes subtle, imperceptible changes. By applying video magnification, these minute colour variations can be amplified and used to extract physiological signals, such as heart rate. This research proposes a machine learning pipeline for colour amplification, trained on the real-world ASC-Emotion dataset. The objective is to develop a model that uses a range of data which includes heart rate derived from colour changes in the skin extracted from video amplification process. This model can be deployed in classrooms to predict “meltdown” events in children with Autism when they are feeling overwhelmed, leading to emotional dysregulation.