Computer Vision Systems Based on Artificial Neural Networks
Artificial Neural Networks have been arguably the most important advancement in artificial intelligence. These AI architectures mimic the connections of neurons in the brain and are capable of learning complex tasks.
Neural networks have been the pillar behind the era of deep learning. Although the architecture has been around for a while, it wasn’t until the late ‘90s that AI researchers got it to work quite well, especially for computer vision – problems that involve visual data like images and videos. Today, neural networks are constantly pushing the boundaries of what is possible in computer vision field. They have been especially useful given the advent of the COVID-19 pandemic. Computer vision systems based on neural networks have been aiding healthcare professionals in the diagnosis of COVID-19. Beyond medicine, these systems have also helped create advances in fields like biology and chemistry. In this article, we will take a look at how computer vision systems are assisting with COVID-19 diagnosis and analyze the trend of these systems in a post COVID-19 world.
Analyzing Medical Images with AI
The analysis of medical images has been one of the biggest use cases for neural networks during the COVID-19 pandemic. Radiologists have been overwhelmed with the amount of chest X-rays they need to examine, and AI is playing a key role in speeding up their analysis. Thankfully, many computer vision systems based on neural networks have been developed to alleviate the burden on these radiologists. Care Mentor AI has been working with experts from St. Petersburg State University and the Mariinsky Hospital to create a diagnostic computer vision AI model which diagnoses lung diseases from X-rays and CT scans of the chest. So far, this neural network computer vision system can successfully detect lung pathologies from X-rays with an accuracy of 84%.
Furthermore, COVID-Net, a convolutional neural network designed for the detection of COVID-19 from chest X-rays, has been released and open sourced by researchers at the University of Waterloo. Chinese technology stronghold, Alibaba, has also released a computer vision-based system that can diagnose suspected COVID-19 cases from CT scans. According to Alibaba, the neural network system diagnoses cases in 20 seconds, with 96% accuracy. This system has been used in 26 hospitals and has helped diagnose around 30,000 suspected COVID-19 cases.
AI Systems in Biology and Chemistry
Beyond medical imaging, computer vision systems based on neural networks have been applied in biology and chemistry. Most notably, these systems have been applied in the analysis of 3D protein structures. For example, Google DeepMind recently released a neural network based model that can predict the protein structure of an organism based on genomic data. This AI system can potentially be applicable in determining which drugs could successfully treat COVID-19 patients.
The exciting progress made with computer vision systems based on neural networks is one that has come to stay. It is imperative that companies start looking at ways to harness these solutions and properly position themselves such that they are not left behind.
Nowigence has created an AI solution, Pluaris™, that comprehends textual data, like doctors’ reports with human-like interpretive intelligence. It connects the diagnosis with MRI, CT Scan, Mammograms, and X-rays to train machines. In the end, machines work faster, are more accurate, work 24/7, facilitate remote working, and allow clinics to handle more patients in less time.