A novel deep learning-based bioacoustic approach for identification of look-alike white-eye (Zosterops) species traded in wildlife markets

The songbird trade crisis in East and South East Asia has been fuelled by high demand, driving many species to the brink of extinction. This demand, driven by the desire for songbirds as pets, for singing competitions and for prayer animal release has led to the overexploitation of numerous species and the introduction and spread of invasive alien species and diseases to novel environments. The ability to identify traded species efficiently and accurately is crucial for monitoring bird trade markets, protecting threatened species and enforcing wildlife laws. Citizen scientists can make major contributions to these conservation efforts but may be constrained by difficulties in distinguishing 'look-alike' bird species traded in markets. To address this challenge, we developed a novel deep learning-based Artificial Intelligence (AI) bioacoustic tool to enable citizen scientists to identify bird species traded in markets. To this end, we used three major avian vocalization databases to access bioacoustic data for 15 morphologically similar White-eye (Zosterops) species that are commonly traded in Asian wildlife markets. Specifically, we employed the Inception v3 pre-trained model to classify the 15 White-eye species and ambient sound (i.e. non-bird sound) using 448 bird recordings we obtained. We converted recordings into spectrogram (i.e. image form) and used eight image augmentation methods to enhance the performance of the AI neural network through training and validation. We found that recall, precision and F1 score increased as the amount of data augmentation increased, resulting in up to 91.6% overall accuracy and an F1 score of 88.8% for identifying focal species. Through the application of bioacoustics and deep learning, this approach would enable citizen scientists and law enforcement officials efficiently and accurately to identify prohibited trade in threatened species, making important contributions to conservation.