Faster disease detection with Computer Vision

Technology advances are helping healthcare and life sciences companies to treat patients faster. But what about prevention? Is there a way to detect conditions before they become illness and design personalized interventions to avoid their onset? Apparently, yes, computer vision could be the answer!

Computer vision technology was developed in 1968 and thanks to the rapid evolution of artificial intelligence (AI) and machine learning (ML), it has gone mainstream. Computer vision is the data science field in which deep learning AI algorithms are trained to recognize events and trigger automatic alerts. To process massive and complex data in real time, computer vision relies on high compute capacities enabled by GPUs (graphics processing units). Today, computer vision is a prerequisite for analyzing the massive flow of video and images. In the next two years, 99% of video/image content captured for enterprise purposes will be analyzed by machines rather than humans.
The technology has evolved dramatically over the decade and made inroads in the healthcare and life sciences industries. Many organizations are already exploring this technology to support medical imaging/video interpretation. As artificial intelligence and machine learning advances further, we expect the adoption of computer vision to increase manifold. Here are some examples of how computer vision is helping health and life sciences organizations to increase efficiencies and therefore, deliver better patient care.

Detecting eye diseases

Computer vision can be utilized to save human vision and thanks to the evolution in this technology, it can now identify more than 50 eye diseases accurately . Three-dimensional images or scans of patients’ eyes are taken using optical coherence tomography (OCT) and shared with the computer vision platform. The platform scans thousands of sample images acquired from patients that are labelled with the doctor’s diagnosis. AI utilizes these images to learn and, and once the platform is smart enough, it can recognize a pathological image and even infer eye disease based on the patient history.

Skin cancer and detecting tumors

Using AI, the computer vision tool can diagnose skin abnormalities to screen skin cancer. The tool relies on skin photographs labelled with the correct diagnosis to train a system so that it will be able to classify skin disease accurately. This screening system segments patients in cohorts attending to their risk of having skin cancer. The patients with high risks are referred to the dermatologist/oncologist while the patients with medium-low risk could be managed by the physician to minimize the risks that lead to the development of skin cancer. Similarly, computer vision can be beneficial in detecting tumors early and accurately in part of the body that are inaccessible for a biopsy. Tumors spread rapidly across the body if left undetected, making early detection crucial. If medical professionals could rely on computer vision tools to detect tumors faster, this could make a huge difference in a patient’s survival.

Medical imaging

By applying computer vision tools to study X-rays, ultrasound images, endoscopy videos and magnetic resonance imaging, doctors and surgeons can get a glimpse of internal organs and identify abnormalities quickly. Using computer vision, ultrasound specialists could detect abnormalities that the human eye could miss. For example, computer vision can help diagnose congenital disabilities in fetuses.

The ability to use computer vision technologies coupled with MRI to process and analyse images faster and identify abnormalities that are easily missed by the naked eye opens a new horizon of possibilities for diagnosis. This could be extremely useful in diagnosing aneurysms and blood cloths more quickly and providing just-in-time care. Similarly, combining computer vision and tomography or CT scans can enable doctors to identify lesions and injuries faster, hastening treatment and improving survival rates.

Computer vision has progressed a lot since it began and is still evolving. The only thing that can slow its adoption is the lack of data. Since this technology relies on labelled images and videos to make the system learn, the availability of quality images and videos is critical for its progress. Computer vision is becoming an integral part of the healthcare and life sciences industry and as AI and ML evolve, it will get further entrenched within the industry.

Computer vision is becoming an integral part of the healthcare and life sciences industry and as AI and ML evolve, it will get further entrenched within the industry.

Share this blog article

  • Share on Linked In

About Natalia Jiménez Lozano
Director, Atos Life Sciences CoE, Distinguished Expert and member of the Scientific Community
Natalia has a MSc in Biochemistry and PhD in Molecular Biology. Working in life sciences and technology since 2012, following 14 years in bioinformatics research, Natalia’s contributions have already been profound and significant. Having pioneered the development and application of the Precision Medicine model and since seen her vision realised in frontline care, in 2020 she achieved her dream of opening the HPC, AI and Quantum Life Sciences Centre of Excellence in Cambridge. She has been instrumental in starting a new chapter for Atos, partners, and customers to accelerate innovation to improve health outcomes and wellbeing of people worldwide.  

Follow or contact Natalia