File size: 5,728 Bytes
b955e6e
 
 
 
 
 
 
 
 
 
ed369b0
 
b955e6e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ed369b0
 
b955e6e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ed369b0
 
b955e6e
 
 
 
 
 
 
 
deb4b8d
b955e6e
 
 
 
6e30a15
b955e6e
d175689
 
b955e6e
 
b90db61
b955e6e
 
 
 
d175689
b955e6e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0a60eea
b955e6e
 
 
 
 
 
 
 
0a60eea
b955e6e
 
 
 
 
 
 
 
 
0a60eea
b955e6e
 
 
 
 
 
 
 
deb4b8d
b955e6e
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
---
configs:
- config_name: ne-america
  data_files:
  - split: train_id
    path: ne-america/train_id.csv
  - split: train_aux
    path: ne-america/train_aux.csv
  - split: val_id
    path: ne-america/val_id.csv
  - split: val_aux
    path: ne-america/val_aux.csv
  - split: test_id
    path: ne-america/test_id.csv
  - split: test_ood_local
    path: ne-america/test_ood_local.csv
  - split: test_ood_non_local
    path: ne-america/test_ood_non-local.csv
  - split: test_ood_non_moth
    path: ne-america/test_ood_non-moth.csv
- config_name: w-europe
  data_files:
  - split: train_id
    path: w-europe/train_id.csv
  - split: train_aux
    path: w-europe/train_aux.csv
  - split: val_id
    path: w-europe/val_id.csv
  - split: val_aux
    path: w-europe/val_aux.csv
  - split: test_id
    path: w-europe/test_id.csv
  - split: test_ood_local
    path: w-europe/test_ood_local.csv
  - split: test_ood_non_local
    path: w-europe/test_ood_non-local.csv
  - split: test_ood_non_moth
    path: w-europe/test_ood_non-moth.csv
- config_name: c-america
  data_files:
  - split: train_id
    path: c-america/train_id.csv
  - split: train_aux
    path: c-america/train_aux.csv
  - split: val_id
    path: c-america/val_id.csv
  - split: val_aux
    path: c-america/val_aux.csv
  - split: test_id
    path: c-america/test_id.csv
  - split: test_ood_local
    path: c-america/test_ood_local.csv
  - split: test_ood_non_local
    path: c-america/test_ood_non-local.csv
  - split: test_ood_non_moth
    path: c-america/test_ood_non-moth.csv
license: cc-by-nc-sa-4.0
pretty_name: Open-Insect
---
## Description

Open-Insect is curated to benchmark open-set recognition of novel species in biodiversity monitoring, with a focus on insects. It is consists of URLs pointing to image files compiled from the [Global Biodiversity Information Facility (GBIF)](https://www.gbif.org/) as well as other metadata such as latitude, longitude, GBIF speciesKey, and label used to train the classifier. Open-Insect partially builds on a subset of [the AMI dataset](https://zenodo.org/records/11358689).

Run this [download script](https://github.com/RolnickLab/Open-Insect/blob/main/download.sh) to download this huggingface dataset and the original images. 

It consists of in-distribution (ID), out-of-distribution (OOD), and auxiliary data (AUX) for three regions: Northeastern North America (NE-America), Western Europe (W-Europe), and Central America (C-America). 

For C-America, Open-Insect also includes 197 images recently collected from Barro Colorado Island (BCI), Panama. It includes 133 species, 59 of which are likely undescribed by science. This OOD-BCI split is to provide a maximally realistic test case for open-set recognition of novel species in biodiversity monitoring. This dataset is hosted at [huggingface](https://huggingface.co/datasets/yuyan-chen/open-insect-bci).


## Dataset Configurations and Splits

This dataset consists of metadata of 3 regions, Northeastern North America (ne-america), Western Europe (w-europe), and Central America (c-america), and links to download the original images. 

### The In-Distribution (ID) Set

For each region, the in-distribution data (`train_id`, `val_id`, and `test_id`) are from [the AMI dataset](https://zenodo.org/records/11358689), which is licensed under the MIT License.


### Out-of-Distribution and Auxiliary Data

The auxiliary and out-of-distribution (OOD) images used in this dataset are sourced from the [Global Biodiversity Information Facility (GBIF)](https://www.gbif.org/). 

#### License Information
These images are provided under various open licenses, including:

- [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)
- [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/)
- [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)
- [CC0 1.0 (Public Domain)](https://creativecommons.org/publicdomain/zero/1.0/)
- [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)
- No Rights Reserved
- [CC BY 3.0](https://creativecommons.org/licenses/by/3.0/)
- [CC BY-SA 3.0](https://creativecommons.org/licenses/by-sa/3.0/)

Users of this dataset should comply with the terms of the individual licenses associated with each image. This dataset primarily contains metadata and does not redistribute the images themselves.


## Dataset Structure

```
Open-Insect
├── c-america
│   ├── test_id.csv
│   ├── test_ood_local.csv
│   ├── test_ood_non-local.csv
│   ├── test_ood_non-moth.csv
│   ├── train_aux.csv
│   ├── train_id.csv
│   ├── val_id.csv
│   └── val_aux.csv
├── ne-america
│   ├── test_id.csv
│   ├── test_ood_local.csv
│   ├── test_ood_non-local.csv
│   ├── test_ood_non-moth.csv
│   ├── train_aux.csv
│   ├── train_id.csv
│   ├── val_id.csv
│   └── val_aux.csv
├── README.md
└── w-europe
    ├── test_id.csv
    ├── test_ood_local.csv
    ├── test_ood_non-local.csv
    ├── test_ood_non-moth.csv
    ├── train_aux.csv
    ├── train_id.csv
    ├── val_id.csv
    └── val_aux.csv
```

## Licensing and Attribution

This dataset partially builds upon data from [the AMI dataset](https://zenodo.org/records/11358689), which is made available under the [MIT License](https://opensource.org/licenses/MIT). We have used a subset of that dataset and included the original license and attribution as required.

You can find the original license in the `third_party_license/LICENSE` file in this repository.

The additional data in this dataset is distributed under the CC-BY-NC-SA-4.0 License.

---