A 8-neighborhood based generalized Hough transform for fragmented graphics recognition

This paper proposes a 8-neighborhood based generalized Hough transform (SGHT) method developed from the existing generalized Hough transform (GHT) method. The difference chiefly shows in the process of generating the Reference-Table and recognizing the graphics. The new method resolves the difficulty in the process of recognizing fragmented graphics from images by conventional GHT, which can be employed not only in the un-fragmented graphics recognition but also in the fragmented graphics recognition with significant precision and rapidity. The effectiveness of ENHT has been proved in our experiments.