The present invention is an Deep Neural Network based technology relating to diagnosis of Anisomelia, also referred to as Leg Length Discrepancy (LLD). This invention is a system and method, which comprises of a diagnosis device referred to as the “LEG-Minder” device that is typically installed in a diagnosis center setting, and diagnoses for LLD on the basis of a neural network model with patient's leg photos or x-rays thereof; and a neural network learning server referred to as the “LEGislator” which is connected to the Internet and performs Deep Neural Network (DNN) learning, on the individual LLD databases generated by a plurality of the “LEG-Minder” device(s). In particular, the present invention relates to a technology in which patient's leg photos (or x-rays) and the corresponding diagnostic result data are acquired in each diagnosis center and then individually uploaded to the LEGislator; and then, on the basis of the uploaded information the LEGislator performs DNN learning to generate an upgraded neural network model, which is then disseminated to the “LEG-Minder” device(s), providing them the latest learnings, which subsequently helps in improving the diagnosis accuracy. This invention optimizes the diagnosis environment of the diagnosis center for Anisomelia.