The research on accurate and intelligent segmentation of knee joint MRI images is of great significance to reduce the work intensity of clinical doctors and nurses. In order to solve the problem that knee joint MRI image segmentation model needs a large number of high-quality tagged images and excessive labeling workload, a semisupervised learning segmentation network model based on 3D scSE-UNet is proposed. The model adopts a self-training semisupervised learning framework and adds a cSE-block+ module on the basis of the 3D UNet model. This module can enhance the effective features of the feature image from two aspects of space and channel, while suppressing irrelevant features and preserving image edge information more completely. In order to solve the problem of rough edge of pseudolabel caused by model segmentation, a fully connected conditional random field is added to refine the edge of pseudolabel in the process of model training. The effectiveness of the model is verified by open source MRNet dataset and OAI dataset. The results show that the proposed model can achieve the segmentation effect of fully supervised learning through a small number of labeled images and effectively reduce the dependence of knee joint MRI image segmentation on expert labeling data.

Background：Lumbar disc herniation (LDH) is a common and frequent disease in orthopedics. It is caused by degenerative changes of the lumbar spine and compression of the lumbar nerves, with the main clinical manifestation being painful involvement of the lumbar region and the legs，which in severe cases affects the patient's quality of life. The disease is treated in a variety of ways with varying degrees of efficacy, LDH is mainly treated conservatively with oral medication or external therapy of Traditional Chinese medicine (TCM) for mild cases, while surgery is required for severe cases. LDH surgical treatment is effective, but there are still some patients whose symptoms are difficult to improve after surgery and who are not suitable for surgical treatment. Pestle needle therapy is a unique external therapy of TCM method, which is non-invasive, non-painful, not easily infected, easy to operate, and easy for patients to accept. This study aimed to design a randomized controlled trial (RCT) to explore the effectiveness and safety of pestle needle in the treatment of LDH. Methods：Sixty patients will be enrolled and randomly divided into one of two groups: the pestle needle group and the celecoxib group. The pestle needle group will be treated with pestle needle at Yaoyangguan bazhen, Heche-mingqiang section, BL23, BL25, GB30, BL40, BL57, BL60 for 30 minutes. 5 days is a course of treatment, with an interval of 2 days between courses, a total of 3 courses of treatment. The pestle needle manipulation is performed to the extent that the skin is flushed during the treatment. The celecoxib group participants will take one celecoxib capsule once a day for the duration of the study period. The Japanese Orthopaedic Association Back Pain Evaluation Questionnaire (JOABPEQ) score, Visual Analog Scale（VAS）score and Oswestry Disability Index（ODI）score were used as treatment outcome assessment indicators, while adverse events (AEs) were recorded. Patients were evaluated for efficacy at baseline, at the end of each course of treatment. Discussion: This study will determine whether pestle needle is more effective and safer than celecoxib in the treatment of patients with LDH. Trial registration The trial has been registered with the Chinese Clinical Trial Register (ChiCTR) under the Registration March ChiCTR2200057684.