Project Plan

Roles and Responsibilities

Hardware and Container Architecture: David Krauthamer

Hardware and Model Training: Robert Plastina

Software and Model Training: Brandon Boutin

Software

An artificial intelligence model will be used to detect the presence of stray cats, and identify if they have been neutered previously (clipped ear identification). This model will run locally on the deployed hardware, and data will be sent to a remote database for review by users. This database will have a frontend component to make viewing data easier.

Hardware

A camera capable of capturing footage during both day and night will be used for the detection of stray cats. When a cat is detected, footage will be captured and classification of the ears will occur. This processing will take place on a Raspberry Pi, which will then send data to a could provider such as Google (GCP) or Microsoft (Azure).

Work Breakdown

Initial Testing 

Recreating the results of the previous group, as well as general cleanup of the codebase.

Comparison Between TensorFlow and PyTorch 

Evaluating whether a switch to TensorFlow from PyTorch is worth the speed and complexity increase,. TensorFlow would allow for the use of Google Coral acceleration hardware which could speed up the model significantly.

Implementation of Clipped Ear Detection

Once a model framework has been chosen, ensure that detection of cats as well as clipped ears is implemented. 

Move Data From Local Storage to the Cloud

The current solution keeps the database locally on the Pi. In theory this shouldn't be difficult to move offsite, but it will require learning a cloud provider and moving the database there. 

Create Data Visualizations Based on Customer Feedback

The customer has requested an additional feature of a way to visualize data trends over time. For example, approximate population sizes and number of neutered cats within. 

Milestone 1 (September 1st - October 11th)

Milestone 2 ( October 12th - November 29th)

Milestone 3 (February 1st - February 28th)

Final Stage (March 1st - April 26th)