Hugging Face's logo

Spaces:
gdussert
/
MCA_Classifier
Sleeping

App Files Community
MCA_Classifier
File size: 2,596 Bytes
b03ea6b
 
 
 
 
 
 
 
 
 
 
 
 
4e4c19f
 
 
a5226c8
4e4c19f
 
 
 
 
 
 
 
 
 
a5226c8
 
 
 
 
 
 
4e4c19f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
title: MCA Classifier
emoji: 🐠
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.37.0
app_file: app.py
pinned: false
license: cc-by-4.0
short_description: Multi-Crop Attention-based Classifier
---

<div align="center">
<h1>Paying Attention to Other Animal Detections Improves Camera Trap Classification Models</h1>

<a href="https://www.biorxiv.org/content/10.1101/2025.07.15.664849v1"><img src="https://img.shields.io/badge/bioRxiv-2025.07.15.664849-b31b1b" alt="arXiv"></a>
<a href='https://huggingface.co/spaces/gdussert/MCA_Classifier'><img src='https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Demo-blue'></a>


**[LBBE](https://lbbe.univ-lyon1.fr/fr)**; **[CEFE](https://www.cefe.cnrs.fr/fr/)**; **[LECA](https://leca.osug.fr/)**


[Gaspard Dussert](https://gdussert.github.io/), [Vincent Miele](https://vmiele.gitlab.io/), [Simon Chamaillé-Jammes](https://simonchamaillejammes.mystrikingly.com/), [Stéphane Dray](https://sdray.github.io/)
</div>

```bibtex
@misc{dussert2025attention,
	author = {Dussert, Gaspard and Dray, Stéphane and Chamaillé-Jammes, Simon and Miele, Vincent},
	title = {Paying Attention to Other Animal Detections Improves Camera Trap Classification Models},
	year = {2025},
	doi = {10.1101/2025.07.15.664849},
    publisher = {bioRxiv},
    type = {preprint},
}
```

## Overview

Multi-Crop Attention-based Classifier uses self-attention to improve classification by levaring the context of other detection within the camera trap sequence. 

GitHub repo is accessible here : [https://github.com/gdussert/MCA_Classifier](https://github.com/gdussert/MCA_Classifier)

Note: Due to the 1GB storage constraint on Hugging Face, the crop_classifier has been converted to float16.

![Gradio Web Interface Preview](images/models.png)

![Gradio Web Interface Preview](images/gradio_screenshot.jpg)


## Acknowledgements

This work was granted access to the HPC resources of IDRIS under the allocation 2022-AD010113729 made by GENCI.

Thanks to [LILA BC](https://lila.science/) for hosting the two datasets, [Snapshot Safari 2024 Expansion](https://lila.science/datasets/snapshot-safari-2024-expansion/) and [Snapshot Serengeti](https://lila.science/datasets/snapshot-serengeti), and for all those who contributed to it.

Thanks [MegaDetector](https://github.com/agentmorris/MegaDetector) and [PytorchWildlife](https://github.com/microsoft/CameraTraps): MDv5a was used for preprocessing and MDV6-Ultralytics-YoloV10-Extra is used by the Gradio demo.

Thanks to the amazing [timm](https://github.com/huggingface/pytorch-image-models) repository.