_ _ (_) ___ | |_ ____ | |/ _ \| __|_ / | | (_) | |_ / / _/ |\___/ \__/___| |__/
Poop Patrol: AI detection of dog fouling
To design, implement and test a machine vision system to automatically detect dog fouling on pavements and other surfaces using a camera.
Dog fouling is a nuisance for most pedestrians, requiring vigilance while walking, and occasionally leading to the unpleasant job of cleaning a dirty shoe or the wheel of a buggy. However, it's a more serious problem for wheelchair users, visually impaired pedestrians and young children who are at increased risk of serious illness caused by exposure to pathogens in dog faeces.
The objective of this project is to design, implement and test a machine vision system that detects dog fouling in images. Possible applications include:
The system will be based on a neural network, possibly implemented using the PyTorch framework. An important element of this project is the creation of a suitable dataset of images, including a large number that contain dog faeces and a large number that don't. All images must be labelled by a human observer, so that the dataset can be used to train, test and validate the neural network. The creation of the dataset can be carried out directly by the student. Alternatively, some form of crowd-sourcing might be used to create a larger set of images.
This project will involve:
No major option is required for this project.