View-based weight network for 3D object recognition

作者：

Highlights：

• Different viewpoint projections of 3D object have different contributions to recognition.

• View-level weights can enhance the performance of network.

• The difference between the first and the second largest probability value can evaluate the results.

• Multiple traditional classifier integrations are effective.

摘要

•Different viewpoint projections of 3D object have different contributions to recognition.•View-level weights can enhance the performance of network.•The difference between the first and the second largest probability value can evaluate the results.•Multiple traditional classifier integrations are effective.

论文关键词：View-based weights layer,3D object recognition,Extreme learning machine,Multi-channel integrated classifier

论文评审过程：Received 16 July 2019, Accepted 5 November 2019, Available online 9 November 2019, Version of Record 21 January 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2019.11.006