What is "Similarity"?
- We score each animal image on how similar it is to other animals. Then we group them.

What is "Confidence?"
- How certain AI-model is in its guess.

By the way, did you know we provide free credits for students & PhD who do academic research?

I showed her Animal Detect with our latest additions and plans. Sara, being sharp minded, challenging, and extremely competent, was the exact feedback I needed to receive on how to improve the platform for the academic world.

The gears are spinning, the furnaces are working.
Stay tuned.

So this is what Animal Detect is aiming to provide. We extended number of data points in export results, and improved the tooling for data annotation and labeling. But this is far from over, there is a big TODO list awaiting to be done. 🥸

The next step is to add an AI-based reading of text contents ON the actual image. The stuff that is written on images by cameras: better timestamps, temperature, moon phase, burst sequences, etc.

Is there anything missing that you need? Tell us. 🙏

You can download now:
- timestamps (when image was taken)
- animal detection label on the image (1 detection = 1 line)
- detection confidence
- full taxonomy: class/order/family/genus/species
- GPS coordinates
- altitude
- image width/height/file size/file format
- camera brand/model etc.

Want to know more about how you can improve your biodiversity monitoring already today? Contact us!

AI didn't recognize your animal? No worries. You can access GBIF database directly on the page and add your animals species from a database with more than 5 million entries. All in less than 5 seconds seconds. The added species will be remembered and available in your list now.

How Animal Detect Works:
1. You upload images
2. System finds all animals and separates blanks.
2. System recognizes 2000+ animal species
3. You get results

Are you part of the team working with camera trap images and want to improve your biodiversity monitoring? Contact me and we can have a chat.

Recently we added a bunch of new features for manual editing like:
- Delete selected images
- GBIF database with 5+ million species entries
- Zoom in/out feature

A bit of facts on the dataset used:
Name: Mega-Efficient Wildlife Classifier (MEWC) Case Study
Size: 660MB
Images: 400 images

The project took me 9 minutes to go through.

A bit of facts on the dataset used:
Name: Mega-Efficient Wildlife Classifier (MEWC) Case Study (unprocessed_service.tar part)
Size: 660MB
Images: 400 images

From 400 images, animals detected on 374 images (93.5%), with 417 total animals & humans detected.

Recently we added a bunch of new features for manual editing like:
- Delete selected images
- GBIF database with 5+ million species entries
- Zoom in/out feature

I believe the nature-credits will be appealing to wildlife teams as long as they do not need to keep an internal office staff to monitor and comply with the reporting policies.

Thanks to NetworkNature EU team for organising this!

Where does Animal Detect stand in the Nature-Credit ecosystem? As a credible report provider. We aim to
help wildlife organisations to convert their wildlife data into results that are audit-friendly, verifiable, and accepted by nature-credit issuing organizations.