Colour Feature Extraction for Computer-Aided Art Therapy

Tadahiko Kimoto (Toyo University, Japan)

Computer analysis of handwritten pastel drawings to extract colour features for use in mental therapy is studied. A measure for evaluating the psychological properties of a colour is defined as the distance to psychological primary colours in a uniform colour space. A colour clustering method suited for representing a drawing region is also described. An experiment was conducted with drawings and mental state values collected in the actual field of clinical therapy. Also, a classifier for predicting mental states from the colour feature vectors was built using a neural network model. After trained with the limited experimental dataset, the classifier proved some effectiveness of the proposed colour features.

Conference: UKSim-AMSS 22nd International Conference on Computer Modelling and Simulation, UKSim2020

Published: Mar 25, 2020

DOI: 10.5013/IJSSST.a.21.02.05