Semi-supervised Learning for RGB-D Object Recognition

CASIA OpenIR > 模式识别实验室

	Semi-supervised Learning for RGB-D Object Recognition
	Yanhua Cheng; Xin Zhao; Kaiqi Huang; Tieniu Tan
	2014
会议名称	International Conference on Pattern Recognition
会议录名称	Proc. International Conference on Pattern Recognition 2014
页码	2377-2382
会议日期	2014-08-01
会议地点	Stockholm, Sweden
摘要	Conventional supervised object recognition methods have been investigated for many years. Despite their successes, there are still two suffering limitations: (1) various information of an object is represented by artificial features only derived from RGB images, (2) lots of manually labeled data is required by supervised learning. To address those limitations, we propose a new semi-supervised learning framework based on RGB and depth (RGB-D) images to improve object recognition. In particular, our framework has two modules: (1) RGB and depth images are represented by convolutional-recursive neural networks to construct high level features, respectively, (2) co-training is exploited to make full use of unlabeled RGB-D instances due to the existing two independent views. Experiments on the standard RGB-D object dataset demonstrate that our method can compete against with other state-of-the-art methods with only 20% labeled data.
关键词	Accuracy cameras feature Extraction object Recognition
语种	英语
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/12684
专题	模式识别实验室
通讯作者	Kaiqi Huang
作者单位	中国科学院自动化研究所
第一作者单位	中国科学院自动化研究所
通讯作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Yanhua Cheng,Xin Zhao,Kaiqi Huang,et al. Semi-supervised Learning for RGB-D Object Recognition[C],2014:2377-2382.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
Semi-supervised Lear（645KB）	会议论文		开放获取	CC BY-NC-SA	浏览下载