Analysis of Radiological Images with Image Processing

Instead of looking for a doctor who can analyze radiological images correctly, people will be able to do the right analysis easily with this program.

This program finds which region it belongs to by first passing the loaded radiological image through the first image processing algorithm. Then, it runs the second image processing algorithm and analyzes whether there is any problem in the uploaded radiological image. If the user allows it, it uses the radiological image to expand its database. It uses image processing technology while performing these operations.

People go from doctor to doctor to have the correct analysis of the radiological images taken. Because doctors do different analyzes, some say there is no problem, some say there is a problem. People do not know which doctor to trust. Meanwhile, if there is a disease, people lose their time, money and maybe their health. Almost everyone who wants to have their radiological image analyzed suffers from this problem.

This solution will affect all people who have radiological images taken, because people will not waste their time, people will not lose their money for nothing. More importantly, if there is a progressive disease and it is overlooked by doctors, the program will detect it, and it will directly affect people's health.

I've heard that everyone around me is suffering from these issues. And when I did a little research, I found out that people lose some limbs and sometimes even their lives because of this problem. I contacted radiologists to develop my solution.

Because it is much easier to explain this work to people when it is in prototype form because you can show it.

I used image processing technology to analyze radiological images. I trained artificial intelligence with radiological images I found on the internet. Then I decided to run the program with two algorithms. First of all, I developed it to find which part of the body the uploaded radiological image belongs to, then compare the database related to the section it found and the uploaded radiological image, so that it can be analyzed. If the program detects a problem that has not been seen before and cannot find the explanation of this problem in the database, I have developed it to warn as an unknown problem.

We aim to implement our solution in all hospitals next year, but it has not been implemented yet, but we aim to implement it in approximately 900 hospitals in Turkey next year.

If this program is used for the interpretation of radiological images in all hospitals, we predict that people will experience a great reduction in limb loss. In order to achieve this goal, we will make people trust this program first. For trust, we first develop this program as an auxiliary program for doctors to use in hospitals. Afterwards, if we prove that the program has detected correctly, we can make people trust this program. In this way, we believe that we can greatly reduce the loss of limbs.

We measure our progress by the number of hospitals where the program is used, and of course by how many analyzes the program does and how many of the analyzes are correct.

In fact, there is not much of an obstacle, but if people do not trust the program, if doctors do not take this program into account and use it, we may be prevented from reaching our goal.
However , we also thought of a system for doctors to use this program . With the points system, the doctor will earn points when he sends feedback according to the accuracy of each radiological image and analysis he uploads, and when these points exceed a certain level, he will be able to use his points as gift certificates.
Other than that, there is no major obstacle.

I have been dealing with image processing for 3 years. Until today, I have done a lot of projects from the internet of things to augmented reality.

-

-

We will use this award primarily to get the public to trust us and then to finance the gift certificates that will be given to doctors.