AI-powered cow detection in complex farm environments
Bibliographic record
Abstract
Animal welfare has become a critical issue in contemporary society, emphasizing our ethical responsibilities toward animals, particularly within livestock farming. In addition, the advent of Artificial Intelligence (AI) technologies, specifically computer vision, offers a innovative approach to monitoring and enhancing animal welfare. Cows, as essential contributors to sustainable agriculture and climate management, being a central part of it. However, existing cow detection algorithms face significant challenges in real-world farming environments, such as complex lighting, occlusions, pose variations and background interference, which hinder accurate and reliable detection. Additionally, the model generalization power is highly desirable as it enables the model to adapt and perform well across different contexts and conditions, beyond its training environment or dataset. This study addresses these challenges in diverse cow dataset composed of six different environments, including indoor and outdoor scenarios. More precisely, we propose a novel detection model that combines YOLOv8 with the CBAM (Convolutional Block Attention Module) and assess its performance against baseline models, including Mask R-CNN, YOLOv5 and YOLOv8. Our findings indicate that while baseline models show promise, their performance degrades in complex real-world conditions, which our approach improves using the CBAM attention module. Overall, YOLOv8-CBAM outperformed YOLOv8 by 2.3% in mAP across all camera types, achieving a precision of 95.2% and an mAP@0.5:0.95 of 82.6%, demonstrating superior generalization and enhanced detection accuracy in complex backgrounds. Thus, the primary contributions of this research are: (1) providing an in-depth analysis of current limitations in cow detection under challenging indoor and outdoor environments, (2) proposing a robust general model that effectively detects cows in complex real-world conditions and (3) evaluating and benchmarking state-of-the-art detection algorithms. Potential application scenarios of the model include automated health monitoring, behavioral analysis and tracking within smart farm management systems, enabling precise detection of individual cows, even in challenging environments. By addressing these critical challenges, this study paves the way for future innovations in AI-driven livestock monitoring, aiming to improve the welfare and management of farm animals while advancing smart agriculture. • Developed YOLOv8-CBAM model for improved cow detection accuracy. • Cow dataset includes six environments, addressing occlusions and lighting issues. • Model generalizes well in both indoor and outdoor scenarios. • YOLOv8-CBAM outperformed YOLOv8 by 2.3% in mAP@0.5:0.95. • Enhanced precision of 95.2% achieved across complex farm conditions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".