Any athlete will tell you that their sport is a numbers game. It’s not only adding points on a scoreboard, though. Slight deviations in timing when exiting the starting blocks, for example, can mean the difference between first and second place in a 100-meter dash. And even a miniscule difference in force could turn a three-point shot into a brick. With such a wide margin of failure, it’s no wonder why coaches keep preaching the fundamentals. But sometimes it’s the variables we can’t see that tip the balance in a competition — a fact that was recognized by researchers at the Universidad de Concepción in Concepción, Chile. Perhaps on a mission to get “swole,” the researchers developed a method that combines videos from thermal cameras with AI-based digital processing to enhance weightlifting training. The hidden variables in question here are the small changes in temperature and body positioning that the method tracks throughout an exercise. The changes in body temperature correlate to muscle activation and areas of strain or fatigue, which can be used to prevent injuries, monitor thermal responses, and quantify physical exercise. Courtesy of iStock.com/SrdicPhoto. While most current methods use “before-and-after” snapshots to track body temperature, they only offer a limited view of the complex dynamics happening inside the body. Instead, the researchers’ method uses data taken from either an inexpensive thermal camera in a smartphone or a high-end thermal device, runs it through data processing algorithms, and then uses Google MediaPipe AI software to identify individuals and their body parts within the images and extract the required information. Additional key points, such as the subject’s joints and the barbell, were then identified and used to calculate the corresponding angles and create a frame-by-frame graph to track positional and temperature changes throughout the exercise. So, can you quantify a heated workout on a budget? Yes, but as predicted, there is a hierarchy of options that unfortunately the inexpensive smartphone app just doesn’t top. The researchers found that the process became more complicated with a conventional camera with some movements — such as curling up into a ball or having a weight obscure their body — too difficult for the inexpensive option to precisely capture. So, unfortunately, the method is not quite ready for a mass consumer release. The researchers, however, are at a stage in which they feel they can branch out into other sports, such as basketball, soccer, and their Paralympic counterparts, as well as refine their algorithm to provide real-time, actionable information to users. In the meantime, gym rats who live in the weight room will have the upper hand in training until further testing is completed; so maybe expect more people to be lifting like Eddie Hall or Brian Shaw next time you head out to pump iron. The research was published in Applied Optics (www.doi.org/10.1364/AO.532763).