Diagnosis of suicidal-depressive emotional state using computer analysis of images posted on social networks

the article discusses the issue of diagnosing suicidal and depressive states of users of social networks by analyzing the color scheme of images and inscriptions placed on them. The color preferences of users in groups of images with suicidal and depressive tendencies were studied. The results of the study largely coincide with the results of other studies in this area. A connection was found between a high content of "negative" colors and a set of words characteristic of depressive behavior. The conducted research shows that a large number of the specified words and the color scheme of the preferred images can be an indicator of a complex psycho-emotional state. The results of the study show that indirect text and graphic signs in statements can serve as an additional tool for their suicidal orientation and, together with other tools, increase the accuracy of diagnosing the possibility of suicide among users of social networks. A set of tools has been created and published to enable such research.

suicidal behavior, social networks, abstract and concrete statements, image analysis, computer vision