Before we get started, let’s look at the background on terminology. The older, classic algorithms that we all know and love are rules based, classical computer vision algorithms. However, when most people say AI now, they are often referring to Deep Learning. Deep Learning, also known as Deep Structured Learning, is a subset of Machine Learning, which in turn is a subset of Artificial Intelligence (AI). This Learning can be supervised, semi-supervised, or unsupervised.
What is Deep Learning in digital pathology? This is a question often asked when exploring the benefits of digital pathology and AI algorithms. By addressing this question, we hope to clear up any confusion. We want to ensure you have a solid understanding of the technology that is being offered and discussed in our industry.
Deep Learning and AI algorithms learn from examples instead of relying on expert-based rules. In doing so, it provides better performance than the old, classical computer vision systems. The algorithms generalize better, make less mistakes, and will not require as much parameterization. The performance is superior and will enable partial automation.
Imagine going from black-and-white movies to color or going from dial-up internet to broadband. That is a huge jump in technology! The difference Deep Learning enables is similar, providing a vast improvement compared to previous algorithms. Deep Learning means less mistakes and improved performance.
Gestalt Diagnostics, in partnership with MindPeak, provides the first Deep Learning AI algorithm integrated within a digital workflow solution in the US being used in routine clinical practice for primary diagnosis. Our first live use customer has implemented our solution as an aide to the pathologist in the interpretation of their breast cases. This is a giant step forward for pathology.
Note - content contributed in conjunction with Felix Faber of MindPeak.
Thank you for joining the Gestalt Community.
Oops, there was an error sending your message.
Please try again later.