Combination of Object Tracking and Object Detection for Animal Recognition

Francis Williams; Ludmila Kuncheva; Juan Rodriguez; Samuel Hennessey

doi:10.1109/IPAS55744.2022.10053017

Combination of Object Tracking and Object Detection for Animal Recognition

Francis Williams, Ludmila Kuncheva, Juan Rodriguez, Samuel Hennessey

School of Computer Science & Engineering

University of Burgos

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both.

Original language	English
Title of host publication	Proc. of the Fifth IEEE International Conference on Image Processing, Applications and Systems (IPAS)
Publisher	IEEE
DOIs	https://doi.org/10.1109/IPAS55744.2022.10053017
Publication status	Published - 6 Dec 2022

Keywords

Animal identification
Bounding boxes
Object tracking
Object detection

Access to Document

10.1109/IPAS55744.2022.10053017

Cite this

@inproceedings{d1b77def41f744ac9952ea20207053a7,

title = "Combination of Object Tracking and Object Detection for Animal Recognition",

abstract = "While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both. ",

keywords = "Animal identification, Bounding boxes, Object tracking, Object detection",

author = "Francis Williams and Ludmila Kuncheva and Juan Rodriguez and Samuel Hennessey",

year = "2022",

month = dec,

day = "6",

doi = "10.1109/IPAS55744.2022.10053017",

language = "English",

booktitle = "Proc. of the Fifth IEEE International Conference on Image Processing, Applications and Systems (IPAS)",

publisher = "IEEE",

}

TY - GEN

T1 - Combination of Object Tracking and Object Detection for Animal Recognition

AU - Williams, Francis

AU - Kuncheva, Ludmila

AU - Rodriguez, Juan

AU - Hennessey, Samuel

PY - 2022/12/6

Y1 - 2022/12/6

N2 - While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both.

AB - While methods for object detection and tracking are well-developed for the purposes of human and vehicle identification, animal identification and re-identification from images and video is lagging behind. There is no clarity as to which object detection methods will work well on animal data. Here we compare two state-of-the art methods which output bounding boxes: the MMDetector and the UniTrack video tracker. Both methods were chosen for their high ranking on benchmark data sets. Using a bespoke pre-annotated database of five videos, we calculated the Average Precision (AP) of the outputs from the two methods. We propose a combination method to fuse the outputs of MMDetection and UniTrack and demonstrate that the proposed method is capable of outperforming both.

KW - Animal identification

KW - Bounding boxes

KW - Object tracking

KW - Object detection

U2 - 10.1109/IPAS55744.2022.10053017

DO - 10.1109/IPAS55744.2022.10053017

M3 - Conference contribution

BT - Proc. of the Fifth IEEE International Conference on Image Processing, Applications and Systems (IPAS)

PB - IEEE

ER -

Combination of Object Tracking and Object Detection for Animal Recognition

Abstract

Keywords

Access to Document

Fingerprint

Cite this