Various methods exist to estimate energy expenditure (heart rate monitors, accelerometers, and oxygen consumption), but all have drawbacks such as attaching devices, battery life, and long testing durations. We are interested in determining whether video-based movement analysis can reliably estimate the energetic cost of walking from only video input. This may improve the throughput, cost, accuracy, and/or fidelity of current energetic cost measurement systems. Additionally, we will test the accuracy of machine learning algorithms to detect body point locations across various skin tones.