On robot indoor scene classification based on descriptor quality and efficiency

作者：

Highlights：

• Generation of optimal descriptors for real-time applications.

• Exploit of well-known techniques: Bag-of-Words and Spatial Pyramid.

• Combination of visual and depth information to obtain robust classifiers.

• Extensive evaluation to find the best parameters for indoor scene classification in robotics.

摘要

•Generation of optimal descriptors for real-time applications.•Exploit of well-known techniques: Bag-of-Words and Spatial Pyramid.•Combination of visual and depth information to obtain robust classifiers.•Extensive evaluation to find the best parameters for indoor scene classification in robotics.

论文关键词：Indoor scenes,Semantic classification,Multi-source classification,Descriptor combination,Spatial pyramid technique

论文评审过程：Received 29 August 2016, Revised 24 February 2017, Accepted 25 February 2017, Available online 28 February 2017, Version of Record 8 March 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.02.040