2023-06-29 07:53:48
The doctors studied the images in detail, marking all areas affected by caries on them, and used these images for training and subsequent validation of the functioning of the neural network.
Developing a network to detect dental caries
A first version of an artificial intelligence system developed by US scientists detected cavities in 79.5% of dental images of patients examined during testing of this system.
Mathematicians and physicians from the United States developed a neural network capable of detecting subtle signs of caries development in its images using open machine vision algorithms.
Also read: Artificial Intelligence: What is its importance and risks to humanity?
The press service of the Association of Dental Scientists (IADR) said that the development of the neural network will simplify the work of dentists.
Do you believe in Tooth Worms? 2010 collaboration with @UMSOD_Dentistry to investigate the “worm-like” cylindrical structures emanating from extracted human teeth. @UMBpatricia @UMBaltimore @adeaweb @IADR @DentalResearch @ADEEorg @thermofisher pic.twitter.com/dzuMHTv784
— Dr. Gary Hack (@ghackdds) August 17, 2019
A statement issued by the association stated: “The algorithm developed by scientists has reached an accuracy level of 79.5% when searching for traces of caries in images of the upper and lower jaws obtained using clinical dental imaging systems, which will allow the use of systems based on these algorithms to automate the work of dentists.” and to develop telemedicine applications for smartphones.
Read also: Artificial intelligence and humans: differences and complementarities
This artificial intelligence system was developed by a group of American mathematicians and doctors led by Chokfu Ogfu, assistant professor at Temple University in Philadelphia, USA, and it is a modified version of the well-known YOLOv7 neural network, which is an open source machine vision system that can be trained to recognize A large number of body types in photos and videos.
In order to solve this problem, Ogfu and colleagues prepared a set of photographs of the upper and lower jaws obtained during the examination of several hundred patients treated in the Department of Dentistry at Temple University.
Doctors studied these images in detail, marking all areas affected by caries on them, and used these images for training and subsequent validation of the functioning of the neural network.
According to the researchers, the first version of the private AI system did a good job with the task assigned to it. It identified signs of the presence or absence of caries in 79.5% of the images, detected more than 83% of the affected teeth on them, and correctly identified the caries boundaries for about 80% of the affected enamel areas.
In the past few years, scientists have made great progress in developing artificial intelligence systems and developed neural networks capable of performing complex tasks and even “thinking” creatively. For example, mathematicians from the United States recently developed an artificial intelligence system that enabled them to identify the effects of skin cancer, and their Russian colleagues developed neural networks that detect epileptic foci and diagnose the condition of coronary vessels.
#network #artificial #intelligence #detect #tooth #decay