Reverse Engineering of Reliefs


Complex decorative reliefs are often added to CAD models in such application areas as sign-making, packaging, and ceramics to make product designs more interesting, more characteristic of the company, or of higher intrinsic value.

New reliefs can be created from 2D artwork using software such as Delcam’s ArtCAM, or can be hand-crafted, typically on a planar surface, and reverse engineered using a 3D scanner. Unfortunately, extraction of the relief from the scan of a physical part is only possible when the relief is lying on a planar surface. As in general such reliefs lie on freeform surfaces, reverse engineering methods are needed to separate the relief from the non-planar background, and to assist in the application of the relief to a different base surface.

This project aims to device mathematical methods and algorithms for the reverse engineering and reapplication of reliefs. The work concerns separating scan points belonging to the relief from the background, identifying single units of repeating relief patterns, fitting the background surface and flatting the relief’s underlying surface.

There are some relief objects as shown below:

A duck relief
Bird relief on a textured background
Cyclic Reliefs
A duck relief
Bird relief on a textured background
Cyclic reliefs



Background Surface Estimation for Reverse Engineering of Reliefs S.-L. Liu, R.R. Martin, F.C. Langbein, P.L. Rosin, International Journal of CAD/CAM, vol. 7, no. 4, 2007. PDF

Reverse engineering of reliefs aims to turn an existing relief superimposed on an underlying surface into a geometric model which may be applied to a different base surface. Steps in this process include segmenting the relief from the background, and describing it as an offset height field relative to the underlying surface. We have previously considered relief segmentation using a geometric snake. Here, we show how to use this initial segmentation to estimate the background surface lying under the relief, which can be used (i) to refine the segmentation and (ii) to express the relief as an offset field.

Our approach fits a B-spline surface patch to the measured background data surrounding the relief, while tension terms ensure this background surface smoothly continues underneath the relief where there are no measured background data points to fit. After making an initial estimate of relief offset height everywhere within the patch, we use a support vector machine to refine the segmentation.

Tests demonstrate that this approach can accurately model the background surface where it underlies the relief, providing more accurate segmentation, as well as relief height field estimation. In particular, this approach provides significant improvements for relief concavities with narrow mouths and can segment reliefs with small internal holes.

Boundary detection

Segmenting Reliefs on Triangle Meshes S.-L. Liu, R. R. Martin, F. C. Langbein, P. L. Rosin Proc. ACM Symp. Solid and Physical Modeling SPM 2006, 7-16, ACM Siggraph, 2006. ISBN 1595933581. PDF

This paper finds the boundary between the background and the relief using an adaptive snake. It starts at a simple user-drawn contour, and is driven inwards by a collapsing force until it matches the relief’s boundary. Our method is insensitive to the choice of the initial contour. The snake’s limiting position is controlled by a feature energy term designed to find a step. A refinement strategy is then used to drive the snake into concavities of the relief contour.

Isolatting the relief from a textured background

Segmenting Geometric Reliefs from Textured Background Surfaces S.-L. Liu, R. R. Martin, F. C. Langbein, P. L. Rosin Computer-Aided Design & Applications 4 (5), 565-583, 2007. ISSN 1686-4360. PDF

Segmentation of geometric reliefs from a textured background has various applications in reverse engineering. We consider two approaches to solve this problem. The first classifies parts of a surface mesh as relief or background, and then uses a snake which moves inwards towards the desired relief boundary, which is coarsely located using an energy based on the classification. The second approach initially smoothes the surface to eliminate the background texture, and locates the snake at the relief boundary using an energy based on the step between the background and the relief. We use example scanned models to demonstrate that both approaches are useful, but are suitable for different types of model.

More isolatting the relief from a textured background

Segmenting Periodic Reliefs on Triangle Meshes S. Liu, R. R. Martin, F. C. Langbein, P. L. Rosin In: Maths of Surfaces XII, Eds. R. R. Martin, M. A. Sabin, J. Winkler, 290-306, Springer LNCS 4647, 2007. ISBN 3540738428. PDF

We first briefly review how we segment the relief from the background surface using our previous work. The rest of the paper then concentrates on how we extract a single repeat unit from the relief. To do so, the user provides two points on one relief boundary which are in approximate correspondence on consecutive repeats of the relief. We first refine the relative locations of these points, and then determine a third corresponding point using relief boundary information. These are used to determine three initial cutting planes across the relief. Then surface registration strategies are utilised to refine the correspondence between adjacent repeat units. Finally, we refine the exact locations of the cutting planes by considering only surface information close to the cutting planes. This allows a repeat unit of the periodic relief to be extracted.




Industrial Partner