.png){kind=icon}
返回
RenderGAN: Generating Realistic Labeled Data
.png){kind=icon}
OA
AI
DOI：10.3389/frobt.2018.00066.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Deep Convolutional Neuronal Networks (DCNNs) are showing remarkableperformance on many computer vision tasks. Due to their large parameter space,they require many labeled samples when trained in a supervised setting. The costsof annotating data manually can render the use of DCNNs infeasible. We present anovel framework called RenderGAN that can generate large amounts of realistic,labeled images by combining a 3D model and the Generative Adversarial Networkframework. In our approach, image augmentations (e.g. lighting, background,and detail) are learned from unlabeled data such that the generated images arestrikingly realistic while preserving the labels known from the 3D model. Weapply the RenderGAN framework to generate images of barcode-like markersthat are attached to honeybees. Training a DCNN on data generated by theRenderGAN yields considerably better performance than training it on variousbaselines.
Keyword:
Generative Adversarial Networks
unsupervised learning
social insects
markers
deep learning
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
