Update README.md
Browse files
README.md
CHANGED
@@ -39,4 +39,76 @@ configs:
|
|
39 |
path: data/train-*
|
40 |
- split: validation
|
41 |
path: data/validation-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
path: data/train-*
|
40 |
- split: validation
|
41 |
path: data/validation-*
|
42 |
+
license: cc-by-4.0
|
43 |
+
task_categories:
|
44 |
+
- audio-classification
|
45 |
+
tags:
|
46 |
+
- biology
|
47 |
+
pretty_name: >-
|
48 |
+
InsectSet459: A large dataset for automatic acoustic identification of insects
|
49 |
+
(Orthoptera and Cicadidae)
|
50 |
---
|
51 |
+
|
52 |
+
# InsectSet459: An Open Dataset of Insect Sounds for Bioacoustic Machine Learning
|
53 |
+
|
54 |
+
## Overview
|
55 |
+
InsectSet459 is a comprehensive dataset of insect sounds designed for developing and testing machine learning algorithms for automatic insect identification. It contains 26,399 audio files from 459 species of Orthoptera (crickets, grasshoppers, katydids) and Cicadidae (cicadas), providing 9.5 days of audio material.
|
56 |
+
|
57 |
+
## Key Features
|
58 |
+
- **459 unique insect species** (310 Orthopteran and 149 Cicada species)
|
59 |
+
- **26,399 audio files** in WAV and MP3 formats
|
60 |
+
- **9.5 days** of total audio duration
|
61 |
+
- **Diverse sample rates** ranging from 8 to 500 kHz
|
62 |
+
- **Worldwide geographic coverage** (heavily focused on Europe and North America)
|
63 |
+
- **Creative Commons licensed** (CC-BY-4.0 or CC0)
|
64 |
+
|
65 |
+
## Dataset Split
|
66 |
+
The dataset is divided into training, validation, and test sets with a 60/20/20 split:
|
67 |
+
- **Training set**: 15,873 files (137.3 hours)
|
68 |
+
- **Validation set**: 5,307 files (46.2 hours)
|
69 |
+
- **Test set**: 5,219 files (43.7 hours)
|
70 |
+
|
71 |
+
> **Note**: For the 2025 BioDCASE data challenge, the test set is being held back until the challenge concludes. The full version will be published as version 1.0 afterward.
|
72 |
+
|
73 |
+
## Data Sources
|
74 |
+
Audio recordings were collected from three primary sources:
|
75 |
+
- **xeno-canto**: Orthoptera recordings
|
76 |
+
- **iNaturalist**: Orthoptera & Cicadidae recordings
|
77 |
+
- **BioAcoustica**: Cicadidae recordings
|
78 |
+
|
79 |
+
## Curation Process
|
80 |
+
- Only research-grade observations from iNaturalist
|
81 |
+
- One file per observation for those with multiple audio attachments
|
82 |
+
- Deduplication based on MD5 checksums
|
83 |
+
- Pooling recordings by username, species, location, date, and time
|
84 |
+
- Selection of only one recording from a one-hour period
|
85 |
+
- Minimum 10 sound examples per species
|
86 |
+
- Stereo files converted to mono
|
87 |
+
- Standardized formats (WAV and MP3)
|
88 |
+
- Long recordings (>2 minutes) automatically trimmed
|
89 |
+
- Species nomenclature unified using COL24.4 taxonomy
|
90 |
+
|
91 |
+
## Use Cases
|
92 |
+
- Development of acoustic insect monitoring systems
|
93 |
+
- Biodiversity assessment through passive acoustic monitoring
|
94 |
+
- Testing new audio representation methods for highly variable frequencies
|
95 |
+
- Ecological research and insect population monitoring
|
96 |
+
|
97 |
+
## Citation
|
98 |
+
|
99 |
+
```bibtex
|
100 |
+
@misc{faiß2025insectset459opendatasetinsect,
|
101 |
+
title={InsectSet459: an open dataset of insect sounds for bioacoustic machine learning},
|
102 |
+
author={Marius Faiß and Burooj Ghani and Dan Stowell},
|
103 |
+
year={2025},
|
104 |
+
eprint={2503.15074},
|
105 |
+
archivePrefix={arXiv},
|
106 |
+
primaryClass={cs.SD},
|
107 |
+
url={https://arxiv.org/abs/2503.15074},
|
108 |
+
}
|
109 |
+
```
|
110 |
+
|
111 |
+
## Creators
|
112 |
+
- Marius Faiß (Max Planck Institute of Animal Behavior, Konstanz, Germany)
|
113 |
+
- Burooj Ghani (Naturalis Biodiversity Centre, Leiden, The Netherlands)
|
114 |
+
- Dan Stowell (Tilburg University, The Netherlands)
|