A novel method for determination of best ordering direction for noisy point clouds

In this paper, we propose a method to determine the best regression axis to order the noisy points for curve reconstruction. The fundamental problem of the curve reconstruction is to order the data in most suitable way. We suggested a histogram based feature to be able to determine the goodness of the order for a regression line. The method developed by using the proposed feature tested on the some synesthetic data with different noise levels. This data was selected due to the eigenvector approach gave wrong results when obtaining regression line. The results showed that the proposed method is encouraging.