Visual memory plays an important role in learning to read, write, spell, or draw. The dysfunction of visual memory can be detected by applying tests, such as Benton Visual Retention Test (Benton Test, BVRT), which is based on displaying ten patterns for ten seconds per each pattern and asking the subject to reproduce the displayed patterns. The purpose of our study was to automate the Benton Test by developing a desktop application. We automated the assessment of hand-drawn geometrical shapes (triangles, circles, rectangles) by applying machine learning and image processing techniques, namely, the ResNet50 network trained for recognition of triangles, circles, rectangles and other shapes, filling in the shape, and determining the type of triangle. The proposed algorithm proved its reliability on subjects with limited ability to draw shapes and was the part of a desktop application which may find its use in screening visual perception and visual memory.
|Title of host publication
|Biocybernetics and Biomedical Engineering – Current Trends and Challenges. Lecture Notes in Networks and Systems
|Published - 09.09.2021
|22nd Polish Conference on Biocybernetics and Biomedical Engineering - Institute of Biocybernetics and Biomedical Engineering PAS, Warsaw, Poland
Duration: 19.09.2023 → 21.09.2023