Based on Artificial Intelligence (AI) and its sub-disciplines like machine learning and deep learning, image recognition technology has become part of our everyday existence.
Whether its the facial recognition feature on your smartphone or Google’s image search capabilities, image recognition technology is all around us.
It is here not only to stay but also expand its applicability in more ambitious fields, such as medical diagnoses.
Not too long ago, doctors relied on their own trained eyes to read medical images like x-rays and CT scans and diagnosed diseases from what they saw.
Although they were accurate most of the times, human error was always a factor.
As a matter of fact, the old system of image analysis is still a part and parcel of healthcare in several developing countries.
Lung cancer is the leading cause of death in China where more than 600,000 people succumb to the dreaded disease every year, primarily due to high levels of air pollution.
With a shortage of radiologists in the country, the possibility of human error due to fatigue and overwork increases that much more, what with radiologists having to analyze hundreds of CT scan images every day.
Realizing that this was a problem that would only multiply with time, Infervision founder Chen Kuan decided to dedicate his deep learning and image recognition expertise to the field of medicine.
“In China there are just 80,000 radiologists who have to work through 1.4 billion radiology scans every year,” Kuan told Bernard Marr – a strategic business & technology advisor to governments and companies – last year.
“By using AI and deep learning, we can augment the work of those doctors. In no way will this technology ever replace doctors – it is intended to eliminate much of the highly repetitive work and empower them to work much faster,” he said.
Infervision tied up with the Szechwan People’s Hospital for a pilot project after which the company began working with several top hospitals in the country.
“So what I saw was that a lot of Chinese people, particularly those living outside big cities, do not get to have any regular medical check-up involving medical imaging,” Kuan told Bernard Marr – an internationally best-selling author, popular keynote speaker, futurist, and a strategic business & technology advisor to governments and companies – last year.
“So they often have to wait until they feel something wrong with their body before they go to a big hospital where it can be diagnosed,” by which time it was often too late to be able to do much about it, he said.
“So what we wanted to do is use deep learning to alleviate this huge problem. If we can use it to learn from the past and assist in diagnosing more accurately, we can help solve the problem,” he told Marr.
Infervision has already completed successful pilot tests of its Head CT Augmented Screening platform and has also been working on incorporating machine learning algorithms in the diagnosis and treatment of strokes.
The company used more than 100,000 annotated image scans to train the algorithms and as more live data keeps adding on, the system will become even more efficient at detecting two main stroke types – hemorrhagic and ischemic.
“X-ray is a very old type of medical check-up – in China, for example, no one had mentioned chest X-ray in academic conferences for more than 15 years,” Kuan told Marr. “Until very recently with the arrival of AI.”
“AI has helped radiologists discover problems they previously weren’t able to see. So we are very proud to see radiologists starting to discuss some very interesting and fantastic cases involving AI,” he said.
When Marr asked Kuan about the way radiologists and other related staff reacted to the new diagnostic technology, considering that it had the potential to replace them, the Infervision CEO said that they were, in fact, rather excited about the whole thing.
“They are very excited”, he told Marr. “Two or three weeks ago there was a congress of Chinese radiologists and there was a lot of excitement about what we can do,” he said.
“They realise that we are helping them with the diagnosis but also helping with treatment plans for patients too,” Kuan added.
At the RSNA (Radiological Society of North America) conference in Chicago this week (Nov 25 to Nov 30), Infervision announced it was expanding its machine learning and deep learning applications to include other chest conditions like cardiac calcification – a relatively common indication of early artery disease, which generally shows up on X-rays or CT scans.
“By adding more scenarios under which our AI works, we are able to offer more help to doctors,” Tech Crunch quoted Kuan as saying.
Doctors can easily identify several diseases from a single scan, something which AI needs to be taught (hence the term machine learning).
However, while a doctor takes 15 to 20 minutes to analyze an image, Infervision’s AI can do it in 30 seconds, complete with a report on the observations and with 20 percent more accuracy.
Currently, the Beijing start-up is associated with as many as 280 hospitals, and counting, with 20 of them outside of the country.
Based on Infervision’s track record, a number of western hospitals have changed their opinion about the company; they had been under the impression that Chinese companies lacked the expertise to deliver state-of-the-art solutions.
“Regardless of their technological capability, Chinese start-ups are blessed with access to mountains of data that no start-ups elsewhere in the world could match. That’s an immediate advantage,” Kuan told TechCrunch.