The eye mask image allows the computer to predict whether a person will have a heart attack right away.
It is often said that the eyes are the windows of the soul. But researchers at Google Inc. see it as an "indicator" of personal health. The technology giant is using deep learning to predict a person's blood pressure, age and smoking status by analyzing photos of the human retina. Google's computers use blood vessels to arrange clues. At the same time, a preliminary study shows that these machines can use this information to predict whether a person is at risk for heart disease.
The latest research relies on convolutional neural networks, a deep learning algorithm that is changing the way biologists analyze images. Scientists are using this method to look for mutations in the genome and predict changes in individual cell layouts. Google's approach is part of a new wave of deep learning applications that make image processing simpler and more versatile, and was recently described in a preprint. This method can even discern biological phenomena that are ignored.
“Before, it was impractical to apply machine learning to many areas of biology,†said Philip Nelson, director of engineering at Google Research in Mountain View, Calif., “Now, you can. But what’s even more exciting is Nowadays, machines can find things that humans might not have seen before."
At the same time, cell biologists from the Allen Cell Science Institute in Seattle are using convolutional neural networks to convert gray cell-plane images captured by optical microscopy into 3D images that are color-coded for some organs of the cell. This method eliminates the need to color cells. Coloring is a process that takes more time and complex labs and can destroy cells. In December last year, a team published details of an advanced technology. The technology uses only a few data, such as cell contours, to predict the shape and location of more cellular components.
“What you see today is that machine learning is undergoing an unprecedented change in completing the imaging-related biological tasks,†said Anne Carpenter, head of imaging at the Broad Institute of Harvard University's Broad Institute. In 2015, her interdisciplinary team began using convolutional neural networks to process cell imaging. Today, this network can handle about 15% of the image data in the center. According to Carpenter's prediction, this method will become the main processing mode of the center within a few years. 
Guangdong Kaihua Electric Appliance Co., Ltd. , https://www.kaihuacable.com