Estimation of lung standing size with the application of computer vision algorithms
Yu. V. Bunin, Ye.V. Vakulik, Ростислав Николаевич Михайлусов, V. V. Negoduyko, Kirill Smelyakov, O.V. Yasinsky · Experimental and Clinical Medicine · 2020
Evaluation of spiral computed tomography data is important to improve the diagnosis of gunshot wounds and the development of further surgical tactics. The aim of the work is to improve the results of the diagnosis of foreign bodies in the lungs by using computer vision algorithms. Image gradation correction, interval segmentation, threshold segmentation, three-dimensional wave method, principal components method are used as a computer vision device. The use of computer vision algorithm allows to clearly determine the size of the foreign body of the lung with an error of 6.8 to 7.2%, which is important for in-depth diagnosis and development of further surgical tactics. Computed vision techniques increase the detail of foreign bodies in the lungs and have significant prospects for the use of spiral computed tomography for in-depth data processing. Keywords: computer vision, spiral computed tomography, lungs, foreign bodies.