Algorithm Sees More: Russia Develops AI System for Medical Imaging Analysis
Researchers at the Artificial Intelligence Institute of Innopolis University have introduced a prototype AI system designed to analyze medical images. The new platform is called Innovit.
What Innovit Can Do
The system can analyze several types of medical imaging data, including computed tomography (CT), magnetic resonance imaging (MRI), X-rays, ultrasound studies and mammography. The algorithm automatically detects pathological changes in images, highlights areas of potential abnormalities and produces a detailed text description of its findings.
In practice, the system performs many of the same steps as a radiologist. It examines the image, searches for signs of disease, records changes and generates a medical report. At the same time, the algorithm can detect multiple pathologies in a single scan and mark them directly on the image.
The technology is based on the Florence-2 architecture, a foundational computer vision model. To train the neural network, researchers compiled a dataset of more than 100,000 medical images of different types. The dataset includes images from all anatomical regions and covers a broad spectrum of pathologies. This allows the system not only to detect abnormalities but also to explain them in the form of a medical report similar to one written by a physician.
Why Universal AI Systems Matter
The project was developed by a multidisciplinary team at Innopolis University, including machine learning specialists, data engineers and clinical experts. The research was supported by the Science and Technology Fund of the Republic of Tatarstan.
Today, most AI solutions used in healthcare are highly specialized. Algorithms are typically trained to work with a single type of data. For example, some systems analyze only brain CT scans or lung X-rays. These tools are designed to solve specific diagnostic tasks, such as identifying pneumonia, tumors or other conditions.
However, physicians often need to use several different algorithms to perform a full diagnostic evaluation. Each system requires its own maintenance, updates and configuration. Innovit was designed as a universal platform capable of working with multiple types of medical images. This approach consolidates several diagnostic functions, simplifies the physician’s workflow and supports clinical decision-making without replacing the doctor.
Benefits for Patients and Physicians
The development of technologies like Innovit directly affects the speed and quality of medical diagnostics. Physicians gain a tool that helps them review imaging results more quickly while drawing attention to potential abnormalities that could be overlooked during heavy workloads.
Large medical centers process hundreds of medical images every day. Intelligent systems can accelerate the initial analysis of these images and reduce the workload on radiology specialists.
Such technologies are especially important for regional healthcare systems. The algorithm can serve as an additional assistant for physicians and help expand access to high-quality diagnostics.
For patients, this means faster results and earlier detection of dangerous diseases, potentially before they reach advanced or life-threatening stages.
Importance for Russia’s Technology Development
Projects of this scale illustrate how the field of medical artificial intelligence is advancing in Russia. University research centers are increasingly becoming platforms where both scientific discoveries and practical technological solutions are developed.
Medical image analysis is currently considered one of the fastest-growing segments of digital healthcare. Computer vision algorithms allow healthcare providers to process vast volumes of medical data automatically and improve diagnostic accuracy.
Prospects for Deployment and International Collaboration
In the near future, developers plan to expand the system’s training dataset by incorporating more complex clinical cases. Work is also underway to integrate the algorithm with large language models. This would allow the system to take into account a patient’s medical history, results from previous examinations and data from electronic health records. Such integration could significantly improve the accuracy of diagnostic conclusions.
If the system successfully completes testing and implementation stages, it could become part of digital healthcare platforms and be used within radiology information systems and medical imaging archives.
At the same time, solutions of this type are attracting interest beyond Russia. The global market for AI-driven diagnostic technologies is growing rapidly. Universal medical imaging analysis systems could be used in international research projects, telemedicine services and healthcare institutions in many countries.
