Intracranial Tumor Segmentation

Prize

According to the points of each team in the final evaluation, we select the highest three teams for regular awards.

The prize will only be awarded to teams giving in-person presentation at workshop (Taipei, 2022-12-17).

1. Champion:            AI training server (4U rackmount) x 1
2. 1^st Runner-up:      16TB HDD x 1
3. 3^rd-place:              2TB SSD x 1

Data

For some reasons, we were prohibited from providing real patient dataset in an open way. Alternatively, we are releasing open datasets generated by DDPM¹. Our initial test showed that models trained using synthetic images performed just as well as using real images, if not better. Because synthetic data may make AI systems better and even more ethical², ICTS 2022 will adopt our open synthetic datasets.

If you are still interested in the real patient dataset, NTUH requested a formal contract. Please contact us for further information.

• 1. Dorjsembe, Z., Odonchimed, S., & Xiao, F. (2022, April). Three-Dimensional Medical Image Synthesis with Denoising Diffusion Probabilistic Models. In Medical Imaging with Deep Learning.
• 2. Strickland, E. (2022, March 10). Are you still using real data to train your AI? IEEE Spectrum. Retrieved July 10, 2022, from https://spectrum.ieee.org/synthetic-data-ai

Task

The task is segmentation. The participants are expected to develop their method and produce segmentation labels of the brain tumors to treat, by using the training data.

Date

	(All deadlines are for 23:59 UTC)	DOI
2022-7-17	Release of training data – 1
2022-9-17	Release of training data – 2
2022-10-17	Release of training data – 3
2022-11-17	Release of validation images
~2022-12-1	Submission of test results (models)
2022-12-1	Declaration of the leaderboard
2022-12-17	Presentations at workshop
	(All deadlines are for 23:59 UTC)