Application of generalized regression neural network on the prediction of the main girths of young females based on 3D point cloud data

Abstract

The widths and thicknesses of 110 female undergraduates are obtained from the extracted silhouette, the girths are measured by using software Imageware based on 3D point cloud data, which corresponding to input and output units of the optimized generalized regression neural network (GRNN) and independent and dependent variables of the multiple linear regression (MLR) model. The training and test samples are used to establish and test the validity of GRNN and MLR model. The result shows that the GRNN is superior to the MLR.

Keywords:

• Generalized regression neural network
• Multiple linear regression
• Point cloud data
• Girth
• Prediction

Mathematics Subject Classification 2010:

• 70G10