ADAM (Adam: automatic detection challenge on age-related macular degeneration)

ADAM 组织为半天挑战赛，这是在美国爱荷华州爱荷华市举行的 ISBI 2020 会议的卫星活动。 ADAM 挑战的重点是研究和开发与年龄相关性黄斑变性 (AMD) 诊断和 AMD 患者眼底照片中的病变分割相关的算法。挑战的目标是评估和比较用于在视网膜眼底图像的通用数据集上检测 AMD 的自动算法。我们邀请医学图像分析社区参与开发和测试现有的和新颖的自动眼底分类和分割方法。说明：ADAM：年龄相关性黄斑变性的自动检测挑战链接：https://amd.grand-challenge.org 年龄相关性黄斑变性，简称 AMD，是黄斑区的一种退行性疾病。主要好发于45岁以上人群，发病率甚至高于老年人糖尿病视网膜病变。 AMD的病因尚不完全清楚，可能与遗传、慢性光破坏效应和营养障碍等多种因素有关。 AMD分为干性AMD和湿性AMD。干性 AMD（也称为非渗出性 AMD）不是新生血管。它的特点是视网膜色素上皮（RPE）进行性萎缩。晚期可在检眼镜下观察到玻璃疣和大面积萎缩。湿性 AMD（也称为新生血管性或渗出性 AMD）的特征是 RPE 下有活跃的新生血管形成，随后会导致渗出、出血和瘢痕形成，如果不及时治疗，最终会对感光器造成不可逆的损害和快速的视力丧失。 AMD 的早期诊断对治疗和预后至关重要。眼底照片是基本检查之一。当前数据集由 AMD 和非 AMD（近视、正常对照等）照片组成。在这些照片中可以发现 AMD 的典型迹象包括玻璃疣、渗出、出血等。ADAM 挑战有 4 个任务： 任务 1：AMD 和非 AMD 眼底图像的分类。任务 2：视盘的检测和分割。任务 3：中央凹的定位。任务 4：从眼底图像中检测和分割病变。

The ADAM Challenge is a half-day satellite event of the ISBI 2020 conference held in Iowa City, Iowa, USA. The ADAM Challenge focuses on researching and developing algorithms related to the diagnosis of age-related macular degeneration (AMD) and the segmentation of lesions in fundus photographs of AMD patients. The goal of this challenge is to evaluate and compare automatic algorithms for detecting AMD on a general dataset of retinal fundus images. We invite the medical image analysis community to participate in developing and testing both existing and novel automatic fundus classification and segmentation methods.

Note: ADAM: Automatic Detection Challenge for Age-Related Macular Degeneration. Link: https://amd.grand-challenge.org

Age-related macular degeneration (AMD for short) is a degenerative disease affecting the macula. It primarily occurs in individuals over 45 years of age, with an incidence rate even higher than that of diabetic retinopathy in the elderly. The exact etiology of AMD remains incompletely understood, and may be associated with multiple factors including genetics, chronic photodamage, and nutritional disorders.

AMD is categorized into dry AMD and wet AMD. Dry AMD (also referred to as non-exudative AMD) does not involve neovascularization. It is characterized by progressive atrophy of the retinal pigment epithelium (RPE). In advanced stages, drusen and large-area atrophy can be visualized via ophthalmoscopy. Wet AMD (also known as neovascular or exudative AMD) is marked by active neovascularization beneath the RPE, which subsequently leads to exudation, hemorrhage, and scar formation. If left untreated, it will ultimately cause irreversible damage to photoreceptors and rapid vision loss.

Early diagnosis of AMD is crucial for treatment and prognosis. Fundus photography is one of the fundamental diagnostic examinations. The current dataset comprises fundus photographs from AMD and non-AMD subjects (including cases of myopia, normal controls, etc.). Typical AMD manifestations observable in these photographs include drusen, exudation, hemorrhage, and other related signs.

The ADAM Challenge consists of four tasks:
Task 1: Classification of AMD and non-AMD fundus images.
Task 2: Detection and segmentation of the optic disc.
Task 3: Localization of the fovea.
Task 4: Detection and segmentation of lesions from fundus images.

搜集汇总

背景与挑战

背景概述

ADAM数据集是一个专注于年龄相关性黄斑变性（AMD）自动检测的挑战赛数据集，发布于2020年，由多个研究机构联合发布。该数据集包含眼底照片，用于评估和比较AMD分类、视盘分割、中央凹定位及病变分割的算法，旨在推动医学图像分析在眼科疾病诊断中的应用。