BriefGPT.xyz
Jun, 2021
自监督预训练在皮肤病变分析中的评估
An Evaluation of Self-Supervised Pre-Training for Skin-Lesion Analysis
HTML
PDF
Levy Chaves, Alceu Bissoto, Eduardo Valle, Sandra Avila
TL;DR
本文比较了三种自监督预训练模型和一个有监督的基线模型，在五个数据集上进行了皮肤病变的诊断。结果表明，自监督预训练模型可以在提高准确性和降低结果的变异性方面与有监督的基线模型相媲美，尤其在数据量少的情况下表现更加稳定和优秀。
Abstract
self-supervised pre-training
appears as an advantageous alternative to supervised pre-trained for
transfer learning
. By synthesizing annotations on pretext tasks, self-supervision allows to pre-train models on la
→