In a paper to be published on November 9, computer scientist at Stanford School of Engineering and pathologist at the Stanford School of Medicine report their collaboration to train the computer to analyze breast cancer microscopic images. Computer analysis is more accurate than those who done by humans.
Their computational Model called pathologists, or C-path, machine learning-based method for automatically analyzing images of cancer networks and predict survival of patients.
Train C-roads, there is this researcher known prognosis is used. Computer pored over the image, tumorous structure measuring range and try to use the structure to predict survival of patients. By comparing the results against the known data, computer model to better predict the survival and adapted gradually know what traits are most important and the cancer care is lacking in predicting survival.
"Basically, computer learning," said Daphne Koller, PhD, Professor of computer science and senior author of the paper.
Medical science has long used three particular features to evaluate the cell, the diversity of the outer core (epithelial) cells of the tumor and the frequency with which these cells divide (a process known as mitosis). Three factors was judged by the spectacle with a microscope and print a qualitatively to stratify patients into three groups of breast cancer that predict survival rates.