This paper investigates an attempt to detect genu valgums in patients. A typical diagnosis of genu valgum is based on the observation of how the alignment of the knees are during standing or if legs have different lengths. This system is an attempt to computationally draw a line between what kind of knee is a “genu valgum” and a “normal” knee. This system would also focus on an approach to increase the accuracy of the model to correctly classify this shape “disorder”. Authors would rely on the creation of our dataset from Physiobank (modified) test it on a live subject to establish the standard of performance.