An explainable deep vision system for animal classification and detection in trail-camera images with automatic post-deployment retraining

作者：

Highlights：

• A deep vision framework is presented for animal detection in trail-camera images.

• Novel strategies are proposed for coping with data drift under realistic conditions.

• Hypothesis testing is employed to address the explainability of the devised system.

摘要

•A deep vision framework is presented for animal detection in trail-camera images.•Novel strategies are proposed for coping with data drift under realistic conditions.•Hypothesis testing is employed to address the explainability of the devised system.

论文关键词：Automatic wildlife monitoring,Animal classification and detection,Data drift and retraining,Model explainability,Convolutional neural networks (CNN),Deep learning

论文评审过程：Received 15 August 2020, Revised 22 January 2021, Accepted 24 January 2021, Available online 29 January 2021, Version of Record 9 February 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.106815