METHOD FOR DIAGNOSING HUMAN FEET CONDITION Russian patent published in 2024 - IPC G16H50/00 A61B5/107 G06T7/12 

FIELD: medicine.

SUBSTANCE: invention relates to medicine, namely to diagnostics of foot condition, and can be used for determining the size of the foot and detecting its deformations, monitoring the dynamics of the development and treatment of the disease, determining the indications for the prescription of the orthopaedic product on the foot and requirements for its design and functional properties, as well as for selecting shoes by size. Disclosed is a method for diagnosing the state of a person’s feet, which involves obtaining a plantogram in the form of an electronic optical image of the feet, assigning classification signs to it in the form of values of the age and sex of the patient, main pre-processing of the plantogram, processing the plantograms by marking the images of the feet with placing diagnostically significant points on them based on the analysis of the geometric and optical characteristics of the image, calculation of plantographic and podometric indicators by coordinates of diagnostically significant points, detection of deformation of feet and determination of its type and degree by deviation of current values of indicators from normative ones, taking into account age and sex of the person being examined, at that, a training representative sample is formed in the form of images of feet of sick and healthy people, pre-marked by medical experts, classified by types of deformations and age-sex characteristics of the examined, after the main pre-processing, the plantogram is subjected to additional pre-processing by means of a software-implemented algorithm in the form of the image contours of the feets, removal of image data outside these contours, dividing the plantogram into two separate images — left and right feet, centring each of them, rotation of the right foot image from right to left relative to its longitudinal axis with preservation of classification signs by age and sex of the patient, and patient's foot image marking is performed by means of software-implemented machine learning model, trained on training set, and consists in determining coordinates of diagnostically significant points of the image based on values of its geometric and optical characteristics according to correlation dependences between coordinates of points and values of image characteristics detected in the training sample.

EFFECT: invention provides reducing the length of the examination and improving the reliability of the results of diagnosing the state of the foot.

1 cl