Image analysis of the tree test using wavelet transform, singular value decomposition, and Fourier transform

Abstract

In this study, we proposed three image analysis methods (wavelet transform, singular value decomposition, and Fourier transform) to evaluate drawings of the tree test quantitatively, and demonstrated the analyses to three images of the tree test drawn by schizophrenic patients. Wavelet analysis suggested that information about the position of drawn trees (direction, depth and width of drawn lines) can be captured. Fourier analysis suggested that information about the direction and depth of drawn lines can be captured. Singular value decomposition suggested that information about the position and direction of drawn lines can be captured. Further research is needed to consider the features of mathematical image analysis in detail, and apply them to analysis of the tree test.