File size: 1,580 Bytes
417a24b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: cc-by-4.0
---

This model was trained by Prakash Palanivelu Rajmohan and Renuka Sharma at [WildObs Australia](https://wildobs.org.au/). The official location of the model files is [the WildObs Hugging Face repository](https://huggingface.co/WildObs/WildObs_QLD_WetTropics).

These files are copied to this Hugging face repo because of the easy integration with [AddaxAI](https://addaxdatascience.com/addaxai).

**Developer**

Prakash Palanivelu Rajmohan and Renuka Sharma

**Owner**

WildObs Australia

**Description**

The wildobs_QLD_WetTropics model for Australian Wet Tropics has been trained on ~30k labelled images from 15 classes. We used the Wildlife Observatory of Australia's tagged image repository for the 'Wet Tropics' rainforests (n = 454 camera deployments, 2,184,664 images, 121 species), and refined this to a balanced dataset of the 15 most common species for CV training and testing. We found that fine-tuning SpeciesNet (https://github.com/google/cameratrapai) delivered the highest performance, often exceeding 95% F1-score. We are in the process of publishing our findings as part of this work.

**Classes**

* Alectura_lathami  
* Bos_taurus  
* Canis_familiaris  
* Casuarius_casuarius  
* Felis_catus  
* Heteromyias_cinereifrons  
* Homo_sapiens  
* Hypsiprymnodon_moschatus  
* Megapodius_reinwardt  
* Orthonyx_spaldingii  
* Perameles_nasuta  
* Sus_scrofa  
* Thylogale_stigmatica  
* Uromys_caudimaculatus  
* Wallabia_bicolor  

**Links**

[Learn more](https://wildobs.org.au/)

[License](https://creativecommons.org/licenses/by-4.0/)