| Title |
Relief Extraction from 3-D Data using Polynomial Regression Analysis |
| Authors |
허의주(Eui-Ju Jeo) ; 최강선(Kang-Sun Choi) |
| DOI |
https://doi.org/10.5573/ieie.2019.56.3.73 |
| Keywords |
Relief Extraction ; 3-D Scanning ; Cultural Heritage ; Mesh Processing |
| Abstract |
In this paper, we propose a method to extract reliefs carved on stone monuments by analyzing 3-D scan data using polynomial regression. Due to the rough surface properties of the monument, conventional methods extract not only reliefs of characters, but also even a lot of noisy dents. The reliefs are usually engraved and shaped like a canal. The cross-sectional shape of this canal can be approximated with a symmetrical quadratic curve. In the proposed method, the regions including reliefs and noises are initially obtained using the conventional methods. The regions are regarded as relief candidate regions. For each candidate region, several cross sections of the region are reconstructed. Then, we evaluate how well the cross section is approximated by a quadratic curve in terms of root mean square error (RMSE). If the average RMSE obtained in the candidate region is large, it is classified as noise and removed. The remaining regions with low average RMSE are extracted as relief. Experimental results show that the proposed method clearly extracts the relief from the uneven rough surface, and the performance of the proposed method is 22% higher than that of the conventional methods in terms of F1 scores. |