CKSP: Cross-species knowledge sharing and preserving for universal animal activity recognition

Deep learning techniques are dominating automated animal activity recognition (AAR) tasks with wearable sensors due to their high performance on large-scale labelled data. However, current deep learning-based AAR models are trained solely on datasets of individual animal species, constraining their applicability in practice and performing poorly when training data are limited. In this study, a one-for-many framework is proposed, dubbed Cross-species Knowledge Sharing and Preserving (CKSP), based on sensor data from diverse animal species. Given the coexistence of generic and species-specific behavioural patterns among different species, a Shared-Preserved Convolution (SPConv) module is designed. This module assigns an individual low-rank convolutional layer to each species for extracting species-specific features and employs a shared full-rank convolutional layer to learn generic features. This enables the CKSP framework to learn inter-species complementarity and alleviates data limitations via increasing data diversity. Considering the training conflict arising from discrepancies in data distributions among species, a Species-specific Batch Normalisation (SBN) module is devised that involves multiple BN layers to separately fit the distributions of different species. To validate CKSP's effectiveness, experiments are performed on three public datasets from horses, sheep, and cattle, respectively. The results show that this approach remarkably boosts the classification performance compared to the baseline method (one-for-one framework) solely trained on individual-species data, with increments of 6.04 %, 2.06 %, and 3.66 % in accuracy, and 10.33 %, 3.67 %, and 7.90 % in F1-score for the horse, sheep, and cattle datasets, respectively. This proves the promising capabilities of the method in leveraging multi-species data to augment classification performance. • Extract species-shared/specific features via a Shared-Preserved Convolution module. • Fit distributions of various species via Species-specific Batch Normalisation. • Propose a universal framework to recognise activities across diverse animal species. • Alleviate the data limitation challenge by learning features across various species. • Outperform the model trained solely on data of individual species.