飞桨学习赛：多模态青光眼分级（GAMMA挑战赛任务一）

Who this competition fits

适合已有飞桨端到端建模经验，并能处理2D眼底彩照与3D OCT体数据的医学图像开发者。也适合想练习多模态青光眼分级、且能以个人身份参赛的学习者。

What this competition gives you

Taken from the competition's official page.

• Data200 paired 2D fundus and 3D OCT images, with 100 training and 100 test pairs.
• TutorialBaseline teaching video.
• Reference paperREFUGE challenge paper for glaucoma assessment from fundus photographs.
• Reference paperAGE challenge paper for angle-closure glaucoma evaluation using OCT.
• Reference paperPaper on deep learning detection of advanced and early glaucoma.

From registration to a first submission

1. 01

   PaddlePaddle 2.0+

2. 02

   眼底彩照处理

3. 03

   3D OCT体数据处理

4. 04

   多模态深度学习

5. 05

   小样本学习

6. 06

   医学图像分类

Before you commit: 核心难点是在训练数据对较少的条件下，联合利用2D与3D两种临床影像完成三分类分级，而“无青光眼”类别仍包含其他眼科疾病，并非普通正常样本。提交时还需同时提交预测结果和原始代码，且模型必须使用PaddlePaddle 2.0及以上版本。