Attention-Based Recurrent Neural Network For Automatic Behavior Laying Hen Recognition

TL;DR

This work addresses automatic recognition of laying-hen vocalizations for health and welfare monitoring by formulating a multi-label classification task and introducing an attention-based RNN. It combines time-domain, frequency-domain, and cepstral features (MFCCs and LFCCs) at the syllable level, utilizing a three-module architecture and a nested loss over master classes to handle semantically close labels. The key result is that time-frequency feature fusion with a single-head attention RNN achieves the highest performance ($F1 = 92.75$), outperforming traditional baselines and demonstrating the practicality of this approach for farm-scale hen monitoring. The study provides a Benin-farm chicken vocalization dataset, highlights the trade-offs between accuracy and computational cost, and proposes a path toward multimodal, scalable welfare monitoring.

Abstract

One of the interests of modern poultry farming is the vocalization of laying hens which contain very useful information on health behavior. This information is used as health and well-being indicators that help breeders better monitor laying hens, which involves early detection of problems for rapid and more effective intervention. In this work, we focus on the sound analysis for the recognition of the types of calls of the laying hens in order to propose a robust system of characterization of their behavior for a better monitoring. To do this, we first collected and annotated laying hen call signals, then designed an optimal acoustic characterization based on the combination of time and frequency domain features. We then used these features to build the multi-label classification models based on recurrent neural network to assign a semantic class to the vocalization that characterize the laying hen behavior. The results show an overall performance with our model based on the combination of time and frequency domain features that obtained the highest F1-score (F1=92.75) with a gain of 17% on the models using the frequency domain features and of 8% on the compared approaches from the litterature.

Figures (2)

• Figure 1: Architecture of the hen vocalization multi-label classification system.
• Figure 2: Accuracies and F1 scores of models