Neural network taught to identify anxiety and aggression in children through their drawings

Scientists from Moscow State University of Psychology and Education have created a neural network that analyzes children's drawings and helps psychologists understand the personality traits of a child. AI can identify signs of anxiety or aggression, which will help specialists to provide assistance in time.

Image source generated by the neural network qwenlm.ai

The authors have developed a system that automatically analyzes drawings of children aged 5 to 14 years. During the psychological test "Drawing a Person," participants must draw a person on a piece of paper using only a simple pencil. This test is widely used in education to study the personality of preschoolers and school students.

Despite the apparent simplicity of this technique, the interpretation of the results is often difficult, associated with the risk of arbitrary interpretation, and requires sufficient experience and high qualifications from educational psychologists. The mobile assistant application created on the basis of our system will minimize errors and speed up the image analysis process.

Marina Sorokova, project manager, Head of the Department of Digital Education at Moscow State University of Psychology and Education

According to the creators' idea, the psychologist will be able to take a picture of the drawing with the camera of a mobile phone, send the image to the system, and receive an analysis in the form of personality characteristics of the subject. This solution will simplify the work of specialists in educational institutions and social services.

In addition, Moscow State University of Psychology and Education plans to integrate the application into the training process for psychologists and teachers.

The main task of the neural network complex is to identify, at the first stage of diagnostics, the personal characteristics of children and adolescents that require attention from specialists: a very high level of anxiety or aggressiveness, a low level of regulation, and others.

Ekaterina Klopotova, Associate Professor of the Faculty of Educational Psychology, Senior Researcher at the Center for Interdisciplinary Research of Modern Childhood at Moscow State University of Psychology and Education

For novice psychologists who do not yet have sufficient practical experience, as well as for complex cases when even an experienced psychologist may feel unsure about the decision made, this is of great importance, Klopotova emphasized.

Earlier in Novosibirsk, scientists created an application that uses computer vision technology in real time to analyzes the condition of the skin and offers recommendations for its care.