MRSpineSeg Challenge: Automated Multi-class Segmentation of Spinal Structures on Volumetric MR Images
Latest news：The complete dataset (including training and test images/GT) is publicly accessible （limited to academic research）. To download the dataset, go to the Data page to sign the dataset use license agreement and send the agreement to the e-mail: firstname.lastname@example.org
1、 Competition name:
The 2nd China Society of Image and Graphics (CSIG) Image and Graphics Technology Challenge: MRSpineSeg Challenge: Automated Multi-class Segmentation of Spinal Structures on Volumetric MR Images. 2、 Purpose：
Degenerative spine diseases (e.g., lumbar disc herniation, spinal stenosis, etc.) have become important diseases affecting the health and quality of life of the elderly and working people. These degenerative spinal diseases often cause changes in the structural morphology and mechanical systems of the spine, resulting in pain, such as lumbar disc herniation, reduced disc height, and nerve compression. Magnetic resonance imaging (MRI) as a non-invasive examination method, it has good soft tissue imaging and no radiation. It is a reliable screening method for degenerative spine diseases. In clinical practice, the treatment of degenerative spinal disorders depends largely on physicians’ experience and lacks accurate quantitative analysis tools. 3D automatic segmentation (MR) images of multiclass spinal structures by MRI are a prerequisite for 3D reconstruction of spinal structures. It can provide quantitative analysis tools for building biomechanical models of the spine, simulating stresses in spinal structures, and assessing the prognosis of different treatment options for degenerative spinal diseases.
This competition aims to gather global developers to explore efficient and accurate 3D automatic segmentation of spinal structure in MR images by using artificial intelligence technology. The spinal structure should be segmented includes 10 vertebrae and 9 intervertebral discs. 3、 Organizer:
4、 Requirements for competition participants： It is open to the whole society. Personnel from colleges and universities, scientific research institutions and enterprises can sign up for the competition. The maximum number of each team is four. Each person can only participate in one team. After team registration, the team information cannot be changed. Note: all personnel who have access to the competition data are prohibited from participating in the competition. Those who have not access to the competition data of Southern Medical University can also participate in the competition. Southern Medical University has the right of final interpretation. 5、 Timeline:
July 1, 2021, 23 examples of unlabeled replay test images will be released, and teams will submit all code, models, results in the specified format and documentation, and the final results need to be reproducible.
July 3, 2021 21:00, deadline for submission of replay results, only one result can be submitted for the replay.
July 13, 2021, the winning teams of this track will be announced, and the champion, first runner-up and third runner-up teams of this track can enter the general track defense.
Grand Prix Final: August 6-8, 2021 at the 11th International Conference on Graphic Graphics in Haikou, China for defense and award presentation.
6、 Task settings With the T2-weighted sagittal MR images and the corresponding physician manual segmentation annotated images provided by this competition, the participants need to train a 3D segmentation model to achieve multi-category 3D automatic segmentation of spinal structures, as shown in Figure 1. There are 20 categories of segmentation targets in the images, including 10 vertebrae, 9 intervertebral discs and background.
The competition is divided into preliminary and final test phase. 172 voxel-level labeled training images and 20 test images will be provided in the preliminary test phase. At the end of the preliminary test phase, teams need to submit all codes, results and documents in the specified format. The organizing committee will identify teams that only rely on manual labeling without algorithmic contributions and terminate the participation of the above teams and directly use the open source baseline output results. teams will be selected from the preliminary test phase according to the ranking of evaluation socres. The final test pase will provide 23 test images. The participating teams need to submit all codes, models, results and documents in the specified format, and the final results need to be reproducible.
In the training stage, public data or private data with image level labels and public data or private data without labels are allowed to be used additionally, but any public or private data with voxel or pixel level labels other than competition data is not allowed to be used. When data other than competition is used, additional data used for submission shall be added when submitting results.
Evaluation score: Average Dice Coefficient. The calculation method is: for each image, calculate the Dice coefficient of each spinal structure included in the gold standard, then average the dice coefficients of these spinal structures to obtain the Dice coefficient of each image, and finally average the dice coefficients of the test image to obtain the final average dice coefficient. The competition provides evaluation score calculation code.
The competition uses the average Dice coefficient as the evaluation index, and its Python code is as follows: evaluate_ demo is an example:
7、 Award The top five places in this competition will receive prizes. Champion: RMB 20000, One Runner up: RMB 15000，One Second runner up: RMB 10000，One Excellence Awards: RMB 2500，Two 8、 Contact us Contact: Shumao Pang , School of Biomedical Engineering, Southern Medical University, China Email: email@example.com 9、 Competition communication group You can join the CSIG spine segmentation competition Tencent QQ group: 854917214