Using computer vision, medical images can be analyzed and conditions diagnosed. The quality of diagnosis offered by Deep Learning is already good enough to serve as an aid to professionals and could be especially useful in geographies with a scarcity of qualified human medical professionals such as doctors, radiologists. Image types are many such as plain X-ray, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, molecular imaging. Practical applications are nearly limitless including detecting tumours, clots, blockages.
This is the industry that affords Deep Learning technology the most direct opportunity to massively transform the quality of human lives.
The possible set of applications is limited only by imagination
The use cases represented here are by no means comprehensive.
We believe that from the collaboration of AI researchers with healthcare and pharma professionals, supported by enlightened regulations - the most beneficial solutions will emerge.
Using DNNs, the process of drug screening can be made virtual, faster and more successful. Drug discovery involves the creation of millions of candidate compounds towards finding viable medicines based on many selection criteria including effectiveness and absence of adverse side effects. The process generates an enormous amount of data and ‘screening’ becomes a challenge. Using DNNs, this process can be accelerated and simplified, as networks filter-in and eliminate candidate molecules while considering a range of filtering criteria.
With the availability of genetic data about individuals, the trend towards personalized medicine will accelerate. For the same condition, there might be multiple therapeutic options available – but which will work best for you? Deep Learning allows the opportunity of using data such as medical history, age, genetic lineage, past conditions, diet, stress levels towards creating a customized prescription. The fact that industry is collecting more data, i.e. observations, about treatment and effect will enable DNNs to ‘learn’ then to recommend the medicine best for an individual patient.