Object instance detection with pruned Alexnet and extended training data

作者：

Highlights：

• A B-PA architecture is proposed with AlexNet-level accuracy but 50x fewer parameters.

• This pre-pruning technique builds a light weighted network without pre-training.

• The data extension strategy efficiently solves the data deficiencies problem.

摘要

•A B-PA architecture is proposed with AlexNet-level accuracy but 50x fewer parameters.•This pre-pruning technique builds a light weighted network without pre-training.•The data extension strategy efficiently solves the data deficiencies problem.

论文关键词：Object instance detection,Pruned Alexnet,Binarized normed gradient,Data extension

论文评审过程：Received 20 March 2018, Revised 2 August 2018, Accepted 27 September 2018, Available online 4 October 2018, Version of Record 11 October 2018.

论文官网地址：https://doi.org/10.1016/j.image.2018.09.013