Explaining away results in accurate and tolerant template matching
作者：
Highlights：
• A new method of detecting image features that is tolerant to changes in appearance.
• Method demonstrated for image features that are pixel intensity values (i.e., templates).
• A novel feature of the method is that templates compete to explain the image data.
• Competition implements form of probabilistic inference known as explaining away.
• Template matching is thus performed via a new, statistically principled, approach.
• State-of-the-art results are produced using standard datasets.
摘要
•A new method of detecting image features that is tolerant to changes in appearance.•Method demonstrated for image features that are pixel intensity values (i.e., templates).•A novel feature of the method is that templates compete to explain the image data.•Competition implements form of probabilistic inference known as explaining away.•Template matching is thus performed via a new, statistically principled, approach.•State-of-the-art results are produced using standard datasets.
论文关键词：Template matching,Feature detection,Image matching,Image registration,Correspondence problem,Multi-view vision
论文评审过程：Received 23 April 2019, Revised 2 March 2020, Accepted 12 March 2020, Available online 14 March 2020, Version of Record 20 March 2020.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107337
