Abstract

Tracking individual animals in a group setting is a exigent task for computer vision and animal science researchers. When the objective is months of uninterrupted tracking and the targeted animals lack discernible differences in their physical characteristics, this task introduces significant challenges. To address these challenges, a probabilistic tracking-by-detection method is proposed. The tracking method uses, as input, visible keypoints of individual animals provided by a fully-convolutional detector. Individual animals are also equipped with ear tags that are used by a classification network to assign unique identification to instances. The fixed cardinality of the targets is leveraged to create a continuous set of tracks and the forward-backward algorithm is used to assign ear-tag identification probabilities to each detected instance. Tracking achieves real-time performance on consumer-grade hardware, in part because it does not rely on complex, costly, graph-based optimizations. A publicly available, human-annotated dataset is introduced to evaluate tracking performance. This dataset contains 15 half-hour long videos of pigs with various ages/sizes, facility environments, and activity levels. Results demonstrate that the proposed method achieves an average precision and recall greater than 95% across the entire dataset. Analysis of the error events reveals environmental conditions and social interactions that are most likely to cause errors in real-world deployments.

Highlights

  • It is necessary to observe animals on an individual level in order to assess their health and wellbeing and ensure efficient production

  • One of the most significant challenges to industry is its reliance upon subjective human observation for assessment, which can be as low as only a few seconds per animal each day [1]

  • This paper introduces a long-term tracking strategy that leverages the high-precision detection outputs provided by [28]

Read more

Summary

Introduction

It is necessary to observe animals on an individual level in order to assess their health and wellbeing and ensure efficient production. The last stage of processing that happens on a per-frame basis is ear tag classification (stage 3 in Figure 1), where small image crops around each detected ear location are converted to probability vectors using a classification network. Key contributions of this work include (1) complimentary methods for detection and classification using convolutional neural networks, (2) a probabilistic framework for merging classification likelihoods to detections, and (3) a publicly available dataset for training and evaluating long-term tracking methods under a variety of challenging situations.

Background
Method
Instance Detection and Part Localization
Fixed-Cardinality Track Interpolation
Visual Marker Classification
Training Details and Evaluation Methodology
Ear Tag Classification
Dataset Description
Performance and Analysis
Results
Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call