飞桨学习赛：PALM病理性近视预测

Who this competition fits

适合已有深度学习图像分类基础、想用 PaddlePaddle 完成医学眼底彩照二分类的个人参赛者。参赛者需要能处理带标签的眼底图像，并在 AI Studio 上训练端到端模型；比赛仅支持个人参赛。

What this competition gives you

There is a baseline you can run as-is — starting from it is usually far faster than building from scratch.

• Data800 labelled fundus images for training and 400 labelled test images
• BaselineSelected community project baseline by jsdbzcm
• BaselineSelected community project baseline by 红白黑
• BaselineSelected community project baseline by 驾长风
• BaselineSelected community project baseline by misaka8080
• BaselineSelected community project baseline by Niki_173
• BaselineSelected community project baseline by 羽小翟
• BaselineSelected community project baseline by 假面骑士Quit
• BaselineSelected community project baseline by yyyokay
• BaselineSelected community project baseline by geoyee
• BaselineSelected community project baseline by 咚咚
• BaselineSelected community project baseline by 伍千柒
• BaselineAdditional community project baseline by jsdbzcm
• BaselineAdditional community project baseline by 红白黑
• BaselinePaddle medical imaging project collection
• Reference paperPaper on pathological myopia classification and lesion segmentation
• Reference paperPaper on pathological myopia detection and optic disc segmentation

From registration to a first submission

1. 01

   PaddlePaddle 2.0+

2. 02

   眼底彩照图像分类

3. 03

   卷积神经网络

4. 04

   二分类概率预测

5. 05

   CSV结果文件生成

6. 06

   AI Studio Notebook

Before you commit: 核心难点是从眼底彩照中区分病理性近视与非病理性近视：标签来自结合临床报告、OCT和视野检查等结果的金标准，而模型输入是图像。还必须按指定文件顺序输出PM风险概率，并提交原始代码与预测结果。