Hey Courtney! I've just sent you an email about coming on Variety Hour to talk about your work :) Looking forward to hearing from you!

Hi @CourtneyShuert 

We support NOAA with AI for individual ID for belugas (but from aerial and from lateral surface too). If some of our techniques can be cross-applied please reach out: jason at wildme dot org

Cheers,

Jason

Hi Kyler,

Did you find anything suitable?

Best,

Melissa

Do you thinks an AI could do that ?

GitHub - naturalsolutions/ecoSecrets: ecoSecrets is a web application which enables users to manage their camera traps data

Wow this is amazing! This is how we integrate Biology and Information Technology. 

Hi Stephanie,

It’s great to hear about your exciting work and that you’ll be attending ICCB 2025 — congratulations on presenting!

The Savannah Tracking team is participating at the exhibition, and we’d love for you to stop by our booth 15.

Given your previous interest in collaring domestic dogs and collecting high-resolution GPS data, it might be a great opportunity for us to connect in person and dive deeper into how our lightweight collar solutions — like those currently deployed on Dingoes — could support your project. We'd be happy to show you how our satellite-enabled collars work, demo our data platforms, and explore a potential fit for your needs.

Lively and informative discussion, I would very much like to contribute if there is some active development work with regards to this. 
I have recent experience with using Model Context Protocol (MCP) to integrate various tools & data repositories with LLMs like Claude. I believe this could be a good idea/path whereby we can do the following:
1. use the images & labels along with any meta-data, chunk/index/store it in vector db
2. integrate with existing data sources available by exposing the data through MCP server

3. Use MCP friendly LLM clients (like Claude) to query, visualize and do other open-ended things leveraging the power of LLM and camera trap data from various sources. 
 

Regards,

Ajay

Hi Jeremy,

I am copying here a reply from Juliane Röder, who is not on the forum here:

The GFBio VAT tool seems to do all the desired analysing and visualising stuff - but it isn't specialized on camera trap data. Instead, it harvests GBIF, GFBio data centers, and environmental data that you can combine in any way you want: https://vat.gfbio.org/#/

6:01

Documentation with some example code and notebooks is here: https://docs.vat.gfbio.org

6:04

A standalone tool for camarea trap data and images is nice, but maybe joining forces and adding some functionality to the VAT tool to better handle camera trap data would be a more sustainable solution...?

The VAT tool can give you some simple statistics of the chosen data (e.g. occurences of wild cats in central Europe in winters 2000-2015), and some simple analyses (wild cat occurences and snow depth). It's focus is on geospatial analyses of occurrence data. They do not have any functionality specific for camera trap data - yet.

My point was, that if you want geospatial analyses and/or additional data to compare to camera trap data, you probably don't have to build a new tool.

I'm not involved with the development of the VAT tool, but I could help with the contact. I know the VAT team.

You can contact her on juliane.roeder@uni-marburg.de.

I did a quick exploration of the VAT tool. It is built quite differently, and my first impression is that it is less accessible; however, as you become more familiar with it, you discover a wide range of options to explore for DwcA occurrences.

I believe it is extremely valuable to also have a tool that works directly with the richest data (Camtrap DP), rather than relying on translated data (Darwin Core), where important details and data structure are lost.

I also think the goals of both tools are different: map-based analysis vs. exploration.

But still usefull to get to know about each other :-)