The radiographic view of the Lumbar spine as anteroposterior (AP) and lateral (LA) taken from the front and side view of the patient should classify in the first step of computer-aided diagnosis; it helps to select an algorithm to localize Lumbar spine using Deep Learning Models in the following steps. However, the inclusion of radiology images into the analysis program may be inaccurate. and thus, it cannot be used as a baseline for subsequent spinal localization analysis. Object detection based on a deep learning approach, therefore, needed to determine the percentage of model confidence before it led to a more accurate analysis of the lumbar spine localization. The accuracy of the radiographic view classification of this experiment was 99.4%, Lumbar spine localization on AP view was 95.6%, and, on LA view, localization was 93.5%.