[ad_1]
MOSCOW (Reuters) – Russian scientists at the Polytechnic University of St. Petersburg have discovered a program that can accurately diagnose cancerous tumors in the lungs in 20 seconds.
The system analyzes the data downloaded from the CT scans of the patient's lungs, taking into account the other analyzes, and after treatment, the program publishes a report on the diagnosis of the diseases.
The project leader, Lev Utkin, said they initially "developed an algorithm to look for 6mm lesions, because the doctors themselves adhered to this tactic, but this system is so well thought out that It detects new tumors of smaller size ". "In honor of the science fiction writer" Isaac Azimov "who invented three laws of robots."
The system can self-learn, a feature based on a neural network able to take into account new data and any task resolved to speed up the process. After the training, the program can reach the result in two seconds.
[ad_2]
Source link
