Update README.md
Browse files
README.md
CHANGED
@@ -16,18 +16,18 @@ tags:
|
|
16 |
We present the BirdSet benchmark that covers a comprehensive range of classification datasets in avian bioacoustics.
|
17 |
We offer a static set of evaluation datasets and a varied collection of training datasets, enabling the application of diverse methodologies.
|
18 |
|
19 |
-
|
|
20 |
-
|
21 |
-
| [PER][1] (Amazon Basin)
|
22 |
-
| [NES][2] (Colombia Costa Rica)
|
23 |
-
| [UHH][3] (Hawaiian Islands)
|
24 |
-
| [HSN][4] (high_sierras)
|
25 |
-
| [NBP][5] (NIPS4BPlus)
|
26 |
-
| [POW][6] (Powdermill Nature)
|
27 |
-
| [SSW][7] (Sapsucker Woods)
|
28 |
-
| [SNE][8] (Sierra Nevada)
|
29 |
-
| [XCM][9] (Xenocanto Subset M)
|
30 |
-
| [XCL][10](Xenocanto
|
31 |
|
32 |
[1]: https://zenodo.org/records/7079124
|
33 |
[2]: https://zenodo.org/records/7525349
|
@@ -40,6 +40,11 @@ We offer a static set of evaluation datasets and a varied collection of training
|
|
40 |
[9]: https://xeno-canto.org/
|
41 |
[10]: https://xeno-canto.org
|
42 |
|
|
|
|
|
|
|
|
|
|
|
43 |
##### Train
|
44 |
- Exclusively using focal audio data from Xeno-Canto (XC) with quality ratings A, B, C and excluding all recordings that are CC-ND.
|
45 |
- Each dataset is tailored for specific target species identified in soundscape files.
|
|
|
16 |
We present the BirdSet benchmark that covers a comprehensive range of classification datasets in avian bioacoustics.
|
17 |
We offer a static set of evaluation datasets and a varied collection of training datasets, enabling the application of diverse methodologies.
|
18 |
|
19 |
+
| | train | test | test_5s | size (GB) | #classes |
|
20 |
+
|--------------------------------|--------:|-----------:|--------:|-----------:|-------------:|
|
21 |
+
| [PER][1] (Amazon Basin) | 16,802 | 14,798 | 15,120 | 10.5 | 132 |
|
22 |
+
| [NES][2] (Colombia Costa Rica) | 16,117 | 6,952 | 24,480 | 14.2 | 89 |
|
23 |
+
| [UHH][3] (Hawaiian Islands) | 3,626 | 59,583 | 36,637 | 4.92 | 25 tr, 27 te |
|
24 |
+
| [HSN][4] (high_sierras) | 5,460 | 10,296 | 12,000 | 5.92 | 21 |
|
25 |
+
| [NBP][5] (NIPS4BPlus) | 24,327 | 5,493 | 563 | 29.9 | 51 |
|
26 |
+
| [POW][6] (Powdermill Nature) | 14,911 | 16,052 | 4,560 | 15.7 | 48 |
|
27 |
+
| [SSW][7] (Sapsucker Woods) | 28,403 | 50,760 | 205,200| 35.2 | 81 |
|
28 |
+
| [SNE][8] (Sierra Nevada) | 19,390 | 20,147 | 23,756 | 20.8 | 56 |
|
29 |
+
| [XCM][9] (Xenocanto Subset M) | 89,798 | x | x | 89.3 | 409 |
|
30 |
+
| [XCL][10](Xenocanto Complete) | 528,434| x | x | 484 | 9,734 |
|
31 |
|
32 |
[1]: https://zenodo.org/records/7079124
|
33 |
[2]: https://zenodo.org/records/7525349
|
|
|
40 |
[9]: https://xeno-canto.org/
|
41 |
[10]: https://xeno-canto.org
|
42 |
|
43 |
+
- We assemble a training dataset for each test dataset that is a subset of a complete XC snapshot. We extract all recordings that have vocalizations of the bird species appearing in the test dataset.
|
44 |
+
- Each sample in the training dataset is a recording may have more than one vocalization of the corresponding bird species.
|
45 |
+
- We omit all recordings from XC that are CC-ND.
|
46 |
+
- Snapshot date of XC: 03/10/2024
|
47 |
+
-
|
48 |
##### Train
|
49 |
- Exclusively using focal audio data from Xeno-Canto (XC) with quality ratings A, B, C and excluding all recordings that are CC-ND.
|
50 |
- Each dataset is tailored for specific target species identified in soundscape files.
|