Learning visual variation for object recognition

作者：

Highlights：

• Visual variation such as pose and lighting, often regarded as nuisance, contains valuable information for object recognition.

• Visual variation learning improves object recognition by guiding a CNN to focus on categorical distinctive features.

• Metric learning can be applied to visual variation to improve object representation learning.

摘要

•Visual variation such as pose and lighting, often regarded as nuisance, contains valuable information for object recognition.•Visual variation learning improves object recognition by guiding a CNN to focus on categorical distinctive features.•Metric learning can be applied to visual variation to improve object representation learning.

论文关键词：Object recognition,Multi-task learning,Convolutional neural network

论文评审过程：Received 12 September 2019, Revised 7 February 2020, Accepted 22 March 2020, Available online 8 April 2020, Version of Record 29 April 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2020.103912