All news

Siberian researchers sharpen precision of medical images

The new technology makes it possible to reduce errors in evaluating diagnostic results by 25%

MOSCOW, December 14. /TASS/. Researchers from the Siberian Federal University (SFU), the Institute of Computational Modeling at the Siberian Branch of the RAS, the Siberian State Aerospace University, and the Krasnoyarsk State Medical University have developed a new algorithm to analyze medical images. The new technology makes it possible to reduce errors in evaluating diagnostic results by 25%. The results of the study were reported in the monograph Computer Vision in Control Systems-4.

Any flaws in reading medical images, for example, obtained with X-rays, magnetic resonance tomography (MRI), ultrasound tomography, customarily originate from the quality of images caused by grainy texture, tiny inclusions, and other noise sources. The researchers invented an approach for treating images, which leads to a notable decrease in the impact of such hindrances.

During the first stage after downloading the image, a special filter complex for treating images including an averaging filter, mean value filter, Gaussian filter, and 2D-refining filter, is applied. Further, using a set of methods, the contours of an object are delineated and they are subjected to digital coding, which the authors of the article developed. Only after going through all these stages of treating medical images, can the data be interpreted.

The authors of the article utilized this coding technology in urology and plastic surgery (hernioplasty - hernia removal) and achieved a 15-25% reduction of errors in diagnostics. According to scientists, the proposed technique is especially effective in urology while diagnosing complex cases of multiple kidney stones. In plastic surgery, the coding allows you to accurately control the variability of tissue textures and forecast their regeneration. The researchers suppose that this method is a promising tool in conducting diagnostics in other domains of medicine where a lot depends on the quality of medical images.