Concepts and Selection

2.2 Concepts

Acceptance Criteria

Success Criteria

The device and software provides a comprehensive solution for feral cat feeders to track the number of neutered and non-neutered cats within their colonies. This data can then be used to assist TNR programs in performing their jobs more efficiently, as well as provide concrete information on the efficacy of TNR programs as a humane solution to feral cat populations.

Out of Scope

Assumptions

Dependencies

Constraints

MVP

The MVP currently consists of a Raspberry Pi 4 connected to an IR camera, with all necessary software running locally. If our assumptions are correct then switching from the PyTorch YOLOv7 model to a TensorFlow based model will provide significant performance increases through the use of specialized acceleration hardware. The improved prototype will consist of this new hardware along with the updated software to utilize the new model. We would also like to demonstrate multiple systems working together using a cloud provider as a backend, however this might require two prototypes, which may not be feasible. 

2.3 Concept Selection

On a Scale of 1 to 5