Feature representation for statistical-learning-based object detection: A review
作者：
Highlights：
• We review the feature representation in statistical learning based object detection.
• We categorize and introduce features based on visual properties.
• The pros/cons on feature properties (e.g., descriptiveness, robustness) are discussed.
• Generic techniques such as dimension reduction and combination are introduced.
• We put some emphasis on future challenges in feature design through this review.
摘要
Highlights•We review the feature representation in statistical learning based object detection.•We categorize and introduce features based on visual properties.•The pros/cons on feature properties (e.g., descriptiveness, robustness) are discussed.•Generic techniques such as dimension reduction and combination are introduced.•We put some emphasis on future challenges in feature design through this review.
论文关键词：Object detection,Feature representation,Feature learning,Dimension reduction,Feature combination
论文评审过程：Received 11 June 2014, Revised 12 February 2015, Accepted 15 April 2015, Available online 24 April 2015, Version of Record 16 July 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.04.018
