News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
Deep learning changes biological analysis image mode
Deep learning changes biological analysis image mode
People often say that eyes are the windows of the mind, but Google researchers regard them as an indicator of people's well-being. Google is using deep learning techniques to predict a person's blood pressure, age, and smoking status through an analysis of people's retinal images. Google's computer can get clues from the distribution of blood vessels, and a previous study indicates that computers can use this information to predict whether a person will have the risk of heart attack in the near future.
These studies are based on a convolution neural network, a depth learning algorithm that can change biologists to analyze images. Scientists are using this method to search for mutations in genes and predict the changes in single cell arrangement. Google has brought a new round of deep learning applications, enabling image processing to be simpler and more general, so that it can identify previously neglected biological phenomena.
Philip Nelson, an engineering executive at Google Research Institute, California mountain city, said: "machine learning has been applied to many areas of biology in the past. As usual, you can do it, and the more exciting is that computers can always observe a lot of details that humans may never have ever seen. "
The convolution neural network can make the computer efficiently and well dispose of the image, and do not need to stop the synthesis of the image. This method was first presented in the field of technology in the 2012, for example, Facebook used this deep learning technique to identify the faces of the photos. But it is difficult for scientists to apply this method to the biological category from time to time, partly because of the cultural differences between the two categories.
Daphne Koller, chief official of San Francisco computer Biologics Corp Calico said: "it is like you have a group of biologists into a team of computer scientists in the room where, they will use different words to talk about each other, and will have a different way of thinking."
Scientists must also be sure what type of research can be discontinued with the use of a convolution neural network. When Google wants to search for mutations in the gene with deep learning, Google scientists must transform the DNA letter chain into an image that the computer can recognize. Then they needed a reference gene to break up the neural network in order to find a mutation. The DeepVariant tool, which came out in December, was able to find small changes in the DNA sequence. In the test, the performance of DeepVariant is at least as traditional tools.
Cell biologists in Seattle Alan Cell Science Institute are making use of convolutional neural network to transform the monotonous gray photos taken by optical microscope into 3D images, and let some cells have color labels. This method eliminates the process of cell staining, which takes more time and needs to be discontinued in the precision laboratory and will cause damage to the cells. Last month, the team released an advanced technology that could predict the shape and location of other parts of the cell with only part of the data.
Massachusetts Institute of Technology Institute of Harvard University and the broad image platform manager Anne Carpenter said: "you always see is a kind of machine learning can There was no parallel in history. changes, with the help of image to complete the task of biology." In 2015, her interdisciplinary team began to use the convolution neural network to dispose of cell images. As usual, Carpenter said that about 15% of the image data at her research center had the aid of the convolution neural network. She predicts that a few years later, this method will be the main image disposal method at the center of the workshop.
What is more exciting is that using convolution neural network to analyze images can unwittingly reveal ingenious biological phenomena, so that biologists begin to ponder over previously neglected problems. Rick Horwitz, executive director of the Aili Institute, said that such an occasional discovery could help medical seminars be progressed from time to time. If deep learning can uncover the ingenious identification of cancer in a single cell, it may help the researchers to identify the tumor early.
Machine learning experts in other biology have been put in the forefront of the field, and the usual convolution neural networks have been widely used in image disposal. "Images are very important, but chemical and molecular data are equally important," said Alex Wolf, a computer biologist at the environmental health center in Germany. I think a serious breakthrough will be completed in the next few years, allowing biologists to apply the convolution neural network more generally. " (passerby)
Torgovnik
These studies are based on a convolution neural network, a depth learning algorithm that can change biologists to analyze images. Scientists are using this method to search for mutations in genes and predict the changes in single cell arrangement. Google has brought a new round of deep learning applications, enabling image processing to be simpler and more general, so that it can identify previously neglected biological phenomena.
Philip Nelson, an engineering executive at Google Research Institute, California mountain city, said: "machine learning has been applied to many areas of biology in the past. As usual, you can do it, and the more exciting is that computers can always observe a lot of details that humans may never have ever seen. "
The convolution neural network can make the computer efficiently and well dispose of the image, and do not need to stop the synthesis of the image. This method was first presented in the field of technology in the 2012, for example, Facebook used this deep learning technique to identify the faces of the photos. But it is difficult for scientists to apply this method to the biological category from time to time, partly because of the cultural differences between the two categories.
Daphne Koller, chief official of San Francisco computer Biologics Corp Calico said: "it is like you have a group of biologists into a team of computer scientists in the room where, they will use different words to talk about each other, and will have a different way of thinking."
Scientists must also be sure what type of research can be discontinued with the use of a convolution neural network. When Google wants to search for mutations in the gene with deep learning, Google scientists must transform the DNA letter chain into an image that the computer can recognize. Then they needed a reference gene to break up the neural network in order to find a mutation. The DeepVariant tool, which came out in December, was able to find small changes in the DNA sequence. In the test, the performance of DeepVariant is at least as traditional tools.
Cell biologists in Seattle Alan Cell Science Institute are making use of convolutional neural network to transform the monotonous gray photos taken by optical microscope into 3D images, and let some cells have color labels. This method eliminates the process of cell staining, which takes more time and needs to be discontinued in the precision laboratory and will cause damage to the cells. Last month, the team released an advanced technology that could predict the shape and location of other parts of the cell with only part of the data.
Massachusetts Institute of Technology Institute of Harvard University and the broad image platform manager Anne Carpenter said: "you always see is a kind of machine learning can There was no parallel in history. changes, with the help of image to complete the task of biology." In 2015, her interdisciplinary team began to use the convolution neural network to dispose of cell images. As usual, Carpenter said that about 15% of the image data at her research center had the aid of the convolution neural network. She predicts that a few years later, this method will be the main image disposal method at the center of the workshop.
What is more exciting is that using convolution neural network to analyze images can unwittingly reveal ingenious biological phenomena, so that biologists begin to ponder over previously neglected problems. Rick Horwitz, executive director of the Aili Institute, said that such an occasional discovery could help medical seminars be progressed from time to time. If deep learning can uncover the ingenious identification of cancer in a single cell, it may help the researchers to identify the tumor early.
Machine learning experts in other biology have been put in the forefront of the field, and the usual convolution neural networks have been widely used in image disposal. "Images are very important, but chemical and molecular data are equally important," said Alex Wolf, a computer biologist at the environmental health center in Germany. I think a serious breakthrough will be completed in the next few years, allowing biologists to apply the convolution neural network more generally. " (passerby)
Torgovnik