Deep Learning Techniques for Fine Motor Skills Assessment in Preschool Children

AUTHOR(S):

Konstantinos Strikas, Apostolos Valiakos, Alkiviadis Tsimpiris, Dimitrios Varsamis, Paraskevi Giagazoglou

TITLE

Deep Learning Techniques for Fine Motor Skills Assessment in Preschool Children

ABSTRACT

Fine motor skills are abilities that involve fine motor control, dexterity and precision usually involving hands and eyes coordination, increasing peoples’ self-reliance and self-esteem in everyday activities. Convolutional Neural Networks (CNN) are considered suitable and can be used to classify images with great accuracy. This study aims to evaluate preschool children fine motor skills, using the proposed MotorSkillsCNN model trained with drawings of Greek pupils in public Kindergarten schools. The training of the proposed CNN model is based on Griffiths II Hand and Eye Coordination Scale. A unique dataset that consists of 884 images of children’s drawings, that represent a man or a woman, is structured at this study and evaluated by experts, shaping the labels of the classes to be used by the proposed model. The results showed that automatic detection of fine motor skills is a hard work but is feasible

KEYWORDS

Convolutional Neural Networks, Deep Learning, Fine motor skills, Griffiths test, Preschool

Cite this paper

Konstantinos Strikas, Apostolos Valiakos, Alkiviadis Tsimpiris, Dimitrios Varsamis, Paraskevi Giagazoglou. (2022) Deep Learning Techniques for Fine Motor Skills Assessment in Preschool Children. International Journal of Education and Learning Systems, 7, 43-49