Mary Paterson

PhD Student

Mary Paterson

AI Analysis of Voice to Aid Laryngeal Cancer Diagnosis

One of the first symptoms of throat cancer can be a change in how someone speaks. But, this can also be a symptom of many other, non-cancer, diseases. We want to look at if artificial intelligence (AI) is able to find the throat cancer from recordings of people speaking.

I focus not only on making AI good at finding throat cancer but also making sure that it’ll work in lots of varied settings. Most of the voice recorings that are used to make AI models are recorded in sound proof rooms but most people don’t have on of those in their house or GP! This means that it’s important that the AI models do not change their mind because of background noise or using different recording devices.