Deep Learning Methods for Animal Counting in Camera Trap Images

被引：2

作者：

Wang, Yizhen ^{[1
]}

Zhang, Yang ^{[1
]}

Feng, Yuan ^{[1
]}

Shang, Yi ^{[1
]}

机构：

[1] Univ Missouri, Dept Elect Engn & Comp Sci, Columbia, MO 65211 USA

来源：

2022 IEEE 34TH INTERNATIONAL CONFERENCE ON TOOLS WITH ARTIFICIAL INTELLIGENCE, ICTAI | 2022年

关键词：

camera traps; animal counting; bounding box ensemble; machine learning; deep learning;

D O I：

10.1109/ICTAI56018.2022.00143

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Camera traps are widely used to monitor the biodiversity and population density of animal species. Camera trap images are usually taken in bursts, and the animal counting problem for a sequence of camera trap images is also an important part of evaluating animal population density. In this paper, two new animal counting methods based on l\ficrosoft MegaDetector V4 have been proposed. FilterDetector uses different filters with bounding box ensemble algorithms to achieve more accurate bounding box detection. DLEDetector is an ensemble method that uses two base deep learning models to correct and enhance the detection result of MegaDetector. Our experimental results in iWildCam 2022 competition test dataset show that both methods outperformed the best method in iWildCam 2021 and the baseline method based on MegaDetector V4 in iWildCam 2022 competition by 9.09% and 6.44%, respectively, and ranked first and third in the competition.

引用

页码：939 / 943

页数：5