Learning Center May 1, 2018

AI in medical diagnostics

Futuristic robotic AI doctor

Will we still need doctors? Computer vision and AI are steering medical diagnostics.

Medical imaging has come a long way since Wilhelm Röntgen developed the first x-ray images in 1895, and now encompasses an array of techniques including radiology, MRI, ultrasound and endoscopy to name just a few. Deep learning and other Artificial Intelligence tools are now being applied to image processing to make diagnostics faster, more accurate and more predictive. Companies such as Avalon AI, whose bold vision is to “accelerate the development of a cure for ageing”, Kheiron Medical Technologies, working in the field of breast screening, and Innersight, supporting surgeons in the construction of patients’ surgical plans, have all brought AI products to market in a bid to lead the field of proactive medical imaging. Recently, Google has also jumped on the medical imaging research bandwagon with its retinal imaging developments. Using cameras running AI programs, data is collected and an algorithm interprets the cardiovascular health of the subject from blood vessels in the eye – and not just their condition today but a prediction for the next five years.

Medical diagnostics brought to the palm of your hand

Computer vision and AI are also maturing in handheld devices. Numerous organisations are shrinking scanning and diagnostic equipment, and combining vision and AI to enable these instruments to monitor, calculate and predict. ThinkSono claim to have created the world’s first software to diagnose Deep Vein Thrombosis (DVT) which utilizes image processing software and neural networks to allow a health professional to make a diagnosis via a portable scanner and their smartphone. Despite delays, Butterfly are due to launch iQ this year – a handheld medical ultrasound scanner which uses deep learning and computer vision to change the face of MRI and ultrasound scanning. Signostics have combined MEMS sensors with machine learning and computer vision algorithms to bring miniaturization to bladder screening.

To relieve pressure on health professionals still further, scans and images can now be processed outside the lab, often by the patient themselves, such the BiliScreen app being developed in the Ubiquitous Computing Lab of the University of Washington. This revolutionary technology is designed for use in the early detection of pancreatic cancer and liver disorders. The app uses a smartphone’s built-in camera to photograph a patient’s eye and computer vision processing extracts the white area, or sclera. A machine learning algorithm identifies the levels of bilirubin present in the sclera; bilirubin build-up presents as yellowing of the skin and sclera (jaundice) and is a key factor in identifying liver and pancreatic ailments. The app is able to identify even slightly raised levels that would go unnoticed by traditional monitoring.

Benefiting the developing world

Other organisations have chosen to focus their efforts on those patients generally less well served by technology. MobileODT is in the early stages of bringing cervical cancer screening to women in poorer parts of the world who have not had a reliable screening infrastructure in place due to the under-resourced and remote facilities available to them. Microscope manufacturer Motic has teamed up with Intellectual Ventures and Bill Gates’ Global Good Fund to distribute its EasyScan GO, a microscope which employs AI algorithms to identify malaria parasites in blood samples. Without the need for a clinician to read the results, patients at risk in developing countries can be identified in about 20 minutes.

So will computers replace doctors?

While the volume and accuracy of medical imaging is undoubtedly being accelerated and improved by computer vision and AI, we can’t see the medical professionals hanging up their latex gloves any time soon. What is likely is that these professionals will be able to focus more of their efforts on prevention and treatment of diseases, and leave more of the routine identification processes to the computers. Which is great news for bringing screening and diagnostics to more people more quickly.

At Active Silicon, we’re already proactive with our medical product development, and have passed customer audits to ISO 13485. Contact us to see how our products could advance your medical imaging.

Click here to read the other episodes in our AI Series, or here to subscribe to our newsletter.

Latest News

See More News
Read More Product News
FireBird 8xCXP-12 frame grabber
June 17, 2025

FireBird 8xCXP-12 CoaXPress Frame Grabber

Designed for bandwidth-intensive multi-CoaXPress camera systems, the new FireBird 8xCXP-12 frame grabber with a PCIe…

Read More Learning Center
3D logo with title text
June 11, 2025

Tech Focus: 3D Imaging

3D imaging is the process of capturing spatial depth and geometric structure of objects. It…

Read More Product News
BlueBird SDI to H.264 IP Adapter, video format converter HD-SDI, 3G-SDI to IP Ethernet
May 27, 2025

BlueBird Adapter converts 3G-SDI video to Ethernet/IP output

The latest addition to our BlueBird family of video adapters, the BlueBird SDI – H.264…

Read More Application
An rov with a camera visible
May 20, 2025

Autofocus-Zoom Cameras for ROV Applications

Camera selection for ROVs is an important and often complex process. Several factors influence which…

Upcoming Events

See More Events
Read More about DSEI 2025 9-12 September 2025
Event details DSEI 2025

DSEI 2025

London, UK

DSEI is a leading Defense & Security Show that connects governments, national armed forces, industry…