Bird Sound Classifier Using Self-Supervised Learning

This repository contains the code and resources for our bird sound classifier, which utilizes self-supervised learning for pretraining on extensive unlabeled soundscape recordings collected from 22 monitoring stations in subtropical montane forests of Taiwan. The model is subsequently fine-tuned to identify 31 bird species native to Taiwan's montane regions.

Key Features

1. Focused on dawn chorus bird song recognition
   The model is designed to classify bird songs during the dawn chorus from soundscape recordings. We trained the model using recordings from this specific period and validated its effectiveness through practical inference tests.

2. Addressing data imbalances and cross-domain challenges
   By incorporating a small portion of open-source datasets and employing augmentation techniques, the model effectively handles data imbalance and domain variation issues.

3. Enhanced robustness with a 'NOTA' category
   To improve adaptability in open-set recognition tasks, a "None of the Above" (NOTA) category is introduced, enabling the model to better handle non-target sounds and background noise.

This classifier is tailored for ecological studies and supports bird monitoring in remote subtropical montane ecosystems. For detailed information, please refer to our publication in Ecological Information.

---

MODEL ARCHITECTURE

The classifier uses a transformer-based architecture with a self-supervised pretraining strategy. Below is the architecture diagram:

Pipeline

Overview of the Model

The model pipeline consists of:

1. Pretraining on large-scale unlabeled soundscape data using a self-supervised learning approach.
2. Fine-tuning on a small, labeled dataset of 31 bird species found in Taiwan's montane forests.
3. Inference on soundscape recordings, focusing on dawn chorus bird songs.

FINE-TUNED BIRD SPECIES

Below is the list of 31 bird species included in the fine-tuning process:

No.	Species ID	Scientific Name	Chinese Name	Common Name
1	AA	Abroscopus albogularis	棕面鶯	Rufous-faced Warbler
2	AC	Arborophila crudigularis	臺灣山鷓鴣	Taiwan Partridge
3	AM	Alcippe morrisonia	繡眼畫眉	Morrison's Fulvetta
4	BG	Brachypteryx goodfellowi	小翼鶇	Taiwan Shortwing
5	BS	Bambusicola sonorivox	臺灣竹雞	Taiwan Bamboo-Partridge
6	CR	Cyanoderma ruficeps	山紅頭	Rufous-capped Babbler
7	DI	Dicaeum ignipectus	紅胸啄花	Fire-breasted Flowerpecker
8	EE	Erythrogenys erythrocnemis	大彎嘴	Black-necklaced Scimitar-Babbler
9	FH	Ficedula hyperythra	黃胸青鶲	Snowy-browed Flycatcher
10	GB	Taenioptynx brodiei	鵂鶹	Collared Owlet
11	HA	Heterophasia auricularis	白耳畫眉	White-eared Sibia
12	HAC	Horornis acanthizoides	深山鶯	Yellowish-bellied Bush Warbler
13	HS	Hierococcyx sparverioides	鷹鵑	Large Hawk-Cuckoo
14	LS	Liocichla steerii	黃胸藪眉	Taiwan Liocichla
15	MH	Machlolophus holsti	黃山雀	Taiwan Yellow Tit
16	ML	Myiomela leucura	白尾鴝	White-tailed Robin
17	NV	Niltava vivida	黃腹琉璃	Taiwan Vivid Niltava
18	PA	Periparus ater	煤山雀	Coal Tit
19	PAL	Pnoepyga albiventer	台灣鷦眉(鱗胸鷦眉)	Scaly-breasted Cupwing
20	PC	Picus canus	綠啄木	Gray-headed Woodpecker
21	PM	Parus monticolus	青背山雀	Green-backed Tit
22	PN	Psilopogon nuchalis	五色鳥	Taiwan Barbet
23	PNI	Pyrrhula nipalensis	褐鷽	Brown Bullfinch
24	PP	Pterorhinus poecilorhynchus	棕噪眉(竹鳥)	Rusty Laughingthrush
25	PS	Pericrocotus solaris	灰喉山椒鳥	Gray-chinned Minivet
26	RG	Regulus goodfellowi	火冠戴菊鳥	Flamecrest
27	SB	Schoeniparus brunneus	頭烏線	Dusky Fulvetta
28	SE	Sitta europaea	茶腹鳾	Eurasian Nuthatch
29	TM	Trochalopteron morrisonianum	臺灣噪眉	White-whiskered Laughingthrush
30	TS	Treron sieboldii	綠鳩	White-bellied Green-Pigeon
31	YB	Yuhina brunneiceps	冠羽畫眉	Taiwan Yuhina

MODEL CHECKPOINTS

The following pre-trained and fine-tuned model checkpoints are available for download:

Checkpoint Name	Dataset	Performance (mAP)	Link
Pre-trained (SSL)	Soundscapes	N/A	Download
Fine-tuned	Taiwan Montane Birds	85.6%	Download

Citation

Wei, Y.C., Chen, W.L., Tuanmu, M.L., Lu, S.S., Shiao, M.T., 2025. Advanced montane bird monitoring using self-supervised learning and transformer on passive acoustic data. Ecological Information.https://doi.org/10.1016/j.ecoinf.2024.102927.