PantheraID

PantheraID is a project focused on contributing to wildlife and environmental conservation through the use of technology and data science.

It was born as a Final Graduation Project, analyzing jaguar movement patterns in the southern region of Costa Rica, with the objective of generating useful information for conservation decision-making.

An individual jaguar identification model was also developed based on their unique rosette patterns. The system, based on neural networks, achieved an overall accuracy of 96.6% and was trained using real data from camera traps provided by an organization dedicated to large cat monitoring.

This project was awarded second place in the Master's in Data Science program at Universidad Complutense de Madrid.

How was it built?

Each jaguar has a unique coat pattern that functions much like a human fingerprint. Taking this differentiating characteristic into account, the challenge was to design a mathematical formula understandable to a computer, using computer vision to study each individual's pattern, assign a label, and distinguish it from others.

Rosettes: This set of rosettes can be seen on the side of Lupo, a jaguar living in the Osa Peninsula.

From videos, frames were extracted and processed for each jaguar, followed by their labeling and the construction of a structured dataset. Then, data analysis techniques were applied, and the information was split into training and testing sets. The model was trained using neural networks, optimizing the extraction of features from the coat pattern—known as rosettes—and iteratively adjusted until achieving an excellent overall accuracy.

Salida de datos: Respuesta del modelo con el nombre del jaguar identificado y la presición.

2241

Processed images

8

Different jaguars

2

Different locations

96,6%

Overall Accuracy

What problem does it solve?

PantheraID significantly reduces the time spent organizing and classifying data—a task that currently falls to a limited number of experts and can easily overwhelm them when data volumes are high. By automating pre-identification, the system offers a scalable solution for large amounts of information, providing direct support to specialists without replacing their clinical or biological judgment, and improving efficiency in collaborative and citizen science projects.

Why is it revolutionary?

Automated solutions for individual jaguar identification remain limited globally, making PantheraID an innovative proposal with a high potential impact on conservation.

What's next?

  • Connecting with other organizations.
  • Feeding the model with more data from different regions and countrie
  • Scaling the infrastructure.
  • Developing a mobile app for field use
  • Implementing pilots with partner organizations.
  • ¿Cómo colaborar?

¿Cómo colaborar?

Field testing with experts.

Data: Sharing camera trap photos to train the model

Validation: Testing the system in the field with experts.

Diffusion: Helping to connect with other organizations.

How can you test it?


cards
Powered by PayPal

← Back

Thank you for your response. ✨

Key aspects of jaguars