飞桨学习赛：PALM眼底彩照视盘探测与分割

Who this competition fits

适合已有 Python 与 PaddlePaddle 2.0+ 使用经验、想练习医学图像语义分割的个人开发者。参赛者需要能处理眼底彩照，并生成与原图对应的像素级视盘掩码；比赛仅支持个人参赛。

What this competition gives you

Taken from the competition's official page.

• Data800 labelled training fundus photographs and 400 labelled test photographs.
• BaselineCommunity AI Studio project by Automate.
• BaselineCommunity AI Studio project by XiaoHu001.
• BaselineCommunity AI Studio project by lulu喵咩叫.
• BaselineCommunity AI Studio project by 一杆梅子酒.
• BaselineCommunity AI Studio project by 羽小翟.
• BaselineCommunity AI Studio project by Niki_173.
• BaselineCommunity AI Studio project by yyyokay.
• BaselineCommunity AI Studio project by JavaRoom.
• BaselineCommunity AI Studio project by 驾长风.
• BaselineCommunity AI Studio project by jsdbzcm.
• BaselineCommunity AI Studio project by 红白黑.
• BaselineCommunity AI Studio project by 吖吖查.
• BaselineCommunity AI Studio project by PangAY.
• BaselineCommunity AI Studio project by 马骏骁.
• BaselineCommunity AI Studio project by 咚咚.
• Reference paperPaper on pathological myopia detection and optic disc segmentation.
• Reference paperPaper on adversarial optic disc and cup segmentation.

From registration to a first submission

1. 01

   PaddlePaddle 2.0+

2. 02

   医学图像语义分割

3. 03

   眼底彩照处理

4. 04

   像素级掩码标注

5. 05

   PNG结果文件生成

6. 06

   ZIP文件提交

Before you commit: 难点在于同时做好视盘是否存在的检测和视盘区域的像素级分割：视盘不完整或缺失时必须输出全背景，而最终结果综合两项任务。还必须严格按原图名称生成指定像素值的 PNG 掩码并打包提交。