Lumbar Spinal Stenosis (LSS) is one of the main causes of chronic low back pain. Chronic low back pain not only reduces the quality of life of people but also can be an important expense item in the country's economy due to the inability of the person to participate in working life and treatment costs. As in other diseases, rapid diagnosis and early treatment of LSS significantly affect the quality of life of the person. Magnetic Resonance (MR) imaging is one of the methods used to diagnose LSS. Diagnosis by interpreting MR images requires serious expertise, and it has been frequently studied by academics in recent years because it is a system that assists the doctor with an objective approach. This field of study is machine learning, which we can call the subbranch of Artificial Intelligence. Deep learning-based machine learning is very successful in processing biomedical images such as MR. In this study, a model that performs 3-dimensional automatic segmentation on T2 sequence Lumbar MR Images is proposed for the diagnosis of LSS. This 3D LSS segmentation study, according to our knowledge, has the feature of being the first in its field and will be an important resource for those who work in this field. In addition, with the proposed model, parts that cannot be fully opened in LSS surgical operations, especially in the nerve roots, can be fully determined beforehand which will ensure that the patient's complaints are completely eliminated after the operation. In MR images, a total of 6 classes were created and segmentation was carried out, including the spinal disc, canal, thecal sac, posterior element, and other regions and background in the image, which are important for LSS. To measure the success of segmentation, the Intersection over Union (IoU) metric was calculated for each class. 3D segmentation success for the validation set in the dataset; Background (IoU = 0.83), Canal (IoU = 0.61), Disc (IoU = 0.91), Other (IoU = 0.97), Posterior element (IoU = 0.82), and Thecal Sac (IoU = 0.81). The 3D automatic segmentation success rates obtained are quite high and show that a Computer Aided Diagnosis system can be created in LSS diagnosis.