CEPHA29: AUTOMATIC CEPHALOMETRIC LANDMARK DETECTION CHALLENGE

News

17 December 2022: Registration page is live now, participants can start registration.

19 December 2022: Training data & evaluation code released to registered participants.

15 January 2023: The deadline for team registration has been extended to January 31, 2023.

23 January 2023: The validation data has been released to registered participants.

Overview

In modern orthodontics, quantitative cephalometric analysis is the most common clinical and research tool that plays an essential role in orthodontic diagnosis and treatment planning. The accurate and precise localization of cephalometric landmarks enables the quantification and classification of anatomical abnormalities, however, the traditional manual way of marking these landmarks is a very tedious job. Endeavours have constantly been made to develop automated cephalometric landmark detection systems but they are inadequate for orthodontic applications. The main reason for this is that the number of publicly available datasets as well as the images provided for training in these datasets are insufficient for an AI model to perform well. To facilitate the development of robust AI solutions for quantitative morphometric analysis, we propose the CEPHA29 Automatic Cephalometric Landmark Detection Challenge. In this regard, we present a benchmark dataset consisting of 1000 cephalometric X-ray images acquired from 7 different radiographic imaging devices. The clinical experts of our team labelled each cephalogram in the dataset with 29 cephalometric landmarks and the CVM stage of the patient. The challenge requires researchers to develop algorithms that (1) automatically localize 29 anatomical landmarks on the cephalogram, and (2) classify the patient’s cephalogram in one of the six CVM stages. The “Train” set along with the ground truth annotations will be released to the participants so that they can design and train their landmark detection methods, whereas the “Validation” set will be provided to the participants without ground truth annotations so that they can evaluate the performance of their proposed method to qualify for the final round. Finally, a “Test” set will be used by the organizers to finalize the best-performing models/methods. Our challenge will not only help derive forward the research and innovation for automatic cephalometric landmark identification but also prove to be the beginning of a new era in the field. For additional information about the challenge, have a look at our preprint.

Schedule

Important Dates

Date	Description
17 December 2022	Registration opens for CEPHA29 Challenge
19 December 2022	Release of training data & evaluation code
23 January 2023	Release of validation data and submission of results (Phase-1) opens
31 January 2023	Registration deadline and Team formation
03 February 2023	Submission deadline of results for evaluation (Phase-1)
08 February 2023	Leaderboard goes live!
12 February 2023	Submission deadline of short paper reporting method.
20 February 2023	Final submission of containerized algorithms starts!
01 March 2023	Submission of IEEE ISBI double column paper
05 March 2023	Decision for paper acceptance

Registration

For registration, each participant must submit a dually signed form at this link. Our organization team will thoroughly review the registration application and respond to the participant within 24 hours. The participants will be provided with further instructions on how to download the dataset and the helping code via email.

Challenge Rules and Code of Conduct

We urge our applicants to carefully read the Code of Conduct before registering for the challenge:

Every challenge participant must follow the scientific code of conduct and maintain professionalism. Any misuse or inappropriate use of material related to the challenge and workshop on or off the website is strictly prohibited. In case of any misconduct, kindly report the organizers immediately.
Each participant must agree to use the provided data following the scope of the DATA USE AND CONFIDENTIALITY AGREEMENT (DUA). Creating multiple accounts to download the data and submit the results is not permitted at all.
It is not allowed to train the network on external data. However, pre-trained models with ImageNet or MS COCO datasets can be used. On the contrary, pre-trained models with any other dataset are not allowed. Similarly, the use of user-generated labels is also not permitted.
All final submissions must be in the form of Docker containers and must adhere to the template provided by the organizers. It is the user’s responsibility to ensure that their algorithm’s Docker container is in good working order during the preliminary test period before submitting it for the final test set.
Since the submission is in the form of Docker containers, only fully automated methods will be accepted for the final submission. It is not possible to submit interactive methods or manual annotations.
Submissions will be graded solely based on their performance on the final test set and following the evaluation criteria.
It is mandatory to write and submit a technical paper explaining the data split, pre-processing, model architectural design, post-processing and any other necessary details related to the proposed solution.
Participants must understand and agree to include the results of their proposed methods in a comprehensive review article. The results and their comparisons will be used in the preparation of a joint publication, with the challenge organizers(s) leading the write-up process.
Users can group and participate as a team. Teams may have several submissions as long as they differ significantly from the methodology described in technical papers. For the final test set, only 1 submission is allowed for each user/team.
The challenge organizing committee reserves the right to remove/eliminate teams at any point if they find any dispute, disrespectful emails, provocations, unnecessary email flood, incorrect leaderboard results and so on.
The decisions of the organizing committee will be considered final and binding on all the matters about this challenge.

Publication Policy

Each participating team should submit a brief manuscript detailing their proposed method, which will be published in the post-workshop proceedings. Following that, authors of appealing discoveries will be invited to further refine/shape their results and potentially also share their insights in a joint publication. This will feature the work of the top performers in detail, as well as the goals, outcomes, and lessons learned from the CEPHA29 challenge, and will be submitted in a top journal. The journal will be defined based on the results of the challenge and feedback from domain experts.

In case of any questions/queries, you may contact us at cepha29.challenge@gmail.com.