Multi-Object Tracking Consistently Improves Wildlife Inference

Mufhumudzi Muthivhi, Jiahao Huo, Fredrik Gustafsson, Terence L. van Zyl · 2026

Read Multi-Object Tracking Consistently Improves Wildlife Inference by Mufhumudzi Muthivhi, Jiahao Huo, Fredrik Gustafsson, Terence L. van Zyl (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-16672.txt.

Description

Muthivhi and colleagues improve wildlife identification from camera traps by linking detections across frames with multi-object tracking. Single-frame classifiers often flip labels for the same animal as noise varies, but tracked trajectories let the method fuse class probabilities into one consensus label. Across three tracking datasets, the approach beats a standalone classifier on every metric. The best tracking models gain 5.1, 3.1, and 2.0 percentage points in weighted F1 score.

Catalog record

CREATORMufhumudzi Muthivhi, Jiahao Huo, Fredrik Gustafsson, Terence L. van Zyl
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26498
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.16672

Browse the full library · About the Tower of Babel