Home  >  All News  >  AI diagnosis Overview

AI diagnosis Overview

Saturday 10th March 2018 Back to list

“Father of computer”  John  von  Neumann predicted that, technology is advancing at an unprecedented remarkable speed. We shall develop towards a singularity. Once we pass this singularity , our familiar society would be totally different. Singularity means origin in the universe dimension, but now it is used to describe a turning point in the social revolution caused by technology break through. We will say, AI (artificial intelligent) is on the singularity of its development.


With its more and more important role in medical industry, AI shall solve many technical and social issues of public health and pave the bring way for human future. While, in some aspects, it brings concerns, to how much extent our work shall be replaced by AI? Will human value decrease?


Here we shall reveal some truth of intelligent diagnosis, which AI already works excellently.

AI works widely in medial diagnosis, one of the most prominent is cancer diagnosis. It is widely known that early and precise diagnosis of cancer makes it possible to become a chronic and controllable disease.


The application of AI diagnosis is as below: 



Working features 

Breast cancer

Google brain and Verily

Locate the position of breast cancer, more precise than professional physicians

Lung cancer

Medical school and Stanford University

Provide suggestions for physicians, it could catch information that would be neglected, high precise diagnosis

Skin cancer 

Medical school and Stanford University

Make precise classification of skin cancer and it could reach the parallel level of professional physicians

1. Locate the position of breast cancer more precisely. Google Science Fair 2012 grants the first glory to a senior student of Wisconsin State. By adopting AI and deep learning technology, on the basis of 7.6 million breast cancer database, he designed a calculation that could enable AI to achieve biopsy precise rate to 96%. This rate is far higher than physicians.  At the same time, the business area advanced to a further step. In 2017, Google, Google brain and Verily company develop an AI product that could diagnose breast cancer with 90% precise rate, far more than human level. 

2. Raise the accuracy rate of lung cancer diagonosis. An earlier experiment indicated that AI could raise lung cancer diagnosis precise rate to 90%. While human physicians only have 50% accuracy rate. AI inspection of tomoscan and X ray inspection are commonly applied. AI could provide suggestions and also catch some part the physicians may neglect. They are more than good assistants but partners of physicians. In 2016, a paper in Nature Report proved that the computer after training could evaluate lung cancer tissue more precisely in lung tissue slicing. They trained the computer software by two thousands lung cancer gene mapping and corresponding databases, and the computer could observe the cancer specificity. 

3. AI could make precise classification of skin cancer and it could reach the parallel level of professional physicians. In 2017, Nature reveals that deep neural network technology could make classification of of skin cancer. The researchers trained the AI network with nearly 130000 pictures of disease. Then they had races between human physicians and machines in classifying skin cancer. The sensibility of AI exceeds 90% and beats all the human physicians. 

Watson from IBM is the representative AI robot in the breakthrough of intelligent diagnosis. IBM has cultivated Watson for years. It is dedicated in cancer diagnosis and is exploring in the diagnosis field of other diseases, for example, diabetes, cardiac disease. Since AI has made the breakthrough of calculation ability, different forms of neural networks have been developed. Image identification is the first success, and AI has challenged human ability in voice identification, art style learning, chess playing in succeed. It plays not assistant but partner of many industries including medical device, finance and service. To mention it, Watson from IBM has started commercial trial in China before being used in scale.