What we have today available technology allows an application that was not possible before.
From what we have of concepts in the market, and there are many, I like the concept of Technopedia that says Artificial Intelligence is the area of computer science that emphasizes the creation of machine intelligence that works and reacts like humans.
The beginning of artificial intelligence systems was using what we have long called “rule engines,” meaning we programmed the computer according to a certain set of rules. (Ex. If the patient has a pain in the upper abdomen, and if on the MRI done, an abnormality in the bowel position appears pointing to an inflammation, then there is a possibility among others that he has a diverticulitis. doctor to make a diagnosis), but far from being a structure of precision and accuracy.
In many existing chatbots today, these models still persist, which causes the famous “artificial stupidity” as I like to call it, since any differences from the rules programming, the system just doesn’t understand, doesn’t process and even understands what is being told to him. Makes you miss the physical attendant.
Today, given the development of the field we have Machine Learning which is the ability of a computer to learn a decision or research model through an algorithm model that with training given to the computer can then have a result where the computer learns while Works on the model.
This area that began with decision tree algorithms, Baynes logic, and others, now uses what we call neural networks, whose focus is to mimic the functioning of the human brain, where all neurons talk to each other.
Within this path, and as I said in the introduction, the technology that enables high capacity and complexity processing in conjunction with high data volumes has allowed the use of more powerful neural networks (neural network clusters with many more contact points) also called Deep learning.
With this, Artificial Intelligence has developed since 2015 in a very powerful way and with training with a very high volume of applications (Evidence) can reach today situations where the computer exceeds the ability of a human being to solve.
In medicine this is no different. However, this does not change the fact that the processing of such applications needs human supervision.
Therefore, these systems are available to support clinical diagnosis, which are then considered clinical decision support systems.
In the area of medical images is also applied Artificial Intelligence that allows analysis of shapes, movement, volume, quantities, anomalous signals, etc. in images and unlike the common Cartesian form that an ordinary human being observes images, the computer analyzes the whole image. even if it finds an anomaly, finding any that might be appearing in the image. Often when we find an anomaly that matches the potential diagnosis, we interrupt the analysis.
The advantages of using Artificial Intelligence in Images can be assessed at 3 different levels:
Finally, the ability to analyze medical images, such as anticipating and detecting a tumor and assessing its type, shape, evolution (follow-up through evolutionary image analysis), or an infection and its range among other possibilities, is greatly enhanced. but when one can also consider in these studies all the patient information, including his demographics, his other imaging studies and other diagnoses in the medical record, and big data medical evidence that allows for a much deeper analysis than a normal human being can do, and using the concepts and algorithms of artificial intelligence can lead to much higher levels of productivity for radiologists, surgeons, medical precision and diagnostics, and a better and more accurate patient experience.