High Accuracy Individual Identification Model of Crested Ibis (Nipponia Nippon) Based on Autoencoder With Self-Attention

Jiangjian Xie,Wenbin Li,Changqing Ding,Jun Yang

doi:10.1109/access.2020.2973243

Abstract

As the population and the distribution of Crested Ibis (Nipponia nippon) become larger, it is necessary to propose a highly efficient census method to estimate the population size of the Crested Ibis. Passive acoustic monitoring (PAM) has a very good prospect for the Crested Ibis monitoring. To realize the automatic census of the Crested Ibis with PAM, the automatic individual identification method based on the vocalization is the key technology. A novel individual identification model was proposed in this paper, which built the autoencoder based on LSTM to obtain the meaningful latent representation from the raw recording directly, further, embedded self-attention and putted forward a combined training mode to achieve distinctive latent representation. With this model, nine Crested Ibis individuals were identified accurately, the highest accuracy is 0.971, and the average accuracy reaches 0.958. As for other three species, Little owl (Athene noctua), Chiffchaff (Phylloscopus collybita) and Tree pipit (Anthus trivialis), the better performances were achieved than the existing method, which means the proposed model can provide an alternative method for the individual identification of other bird species.

Highlights

Crested Ibis (Nipponia nippon) is a globally endangered bird species (IUCN, 2019)
The remainder of this paper is arranged as follows: in Section II, we briefly review the related works on automatic individual identification, autoencoder with Long Short-Term Memory (LSTM) and attention mechanism
Autoencoder based on LSTM had been used for extracting the latent representations of many kinds of data and achieved good performances, such as time sequence data [19], image data [20], [21] and biological data [22] et al C

Summary

INTRODUCTION

Crested Ibis (Nipponia nippon) is a globally endangered bird species (IUCN, 2019). From 1981, many effective conservation measures were taken by the Chinese government, the population size and distribution areas of Crested Ibis have been increasing year by year. The existing census method of the Crested Ibis was proposed [1] In this method, researchers need to count the number of the Crested Ibis individuals in each roosting site at the same time, the summation of each roosting site is the population size of the study area. A novel individual identification method of the Crested Ibis was proposed based on autoencoder and Long Short-Term Memory (LSTM) network. Autoencoder based on LSTM had been used for extracting the latent representations of many kinds of data and achieved good performances, such as time sequence data [19], image data [20], [21] and biological data [22] et al. C.

INDENTIFICATION MODEL OF THE CRESTED IBIS CALL

EXPERIMENTAL RESULTS AND ANALYSIS

Findings

CONCLUSION