Reassembly of fractured objects using surface signature

Developing a Reassembling Algorithm for Broken Objects

“…This is because the method in [14] uses the whole point cloud to compute the final reconstruction, while our method uses a reduced set of keypoints. It can also be observed from the TABLE Ⅱ that the time spent by our method is much less than that by the method in [20]. The reason is because the method in [20] converts the 3D point cloud model into the 2D spin image for description.…”

“…The detailed experimental records of brick pairwise matching are listed in TABLE Ⅱ , where #P is the total number of points in the models after uniform sampling, #F is the number of detected keypoints, init T is the time spent in finding initial matching, mis T is the time in eliminating mismatches, total T is the total time spent in finding the matches. TABLE Ⅱ also compares the time spent by our algorithm and the time by the search algorithms presented in [14] and [20]. It can be seen that our algorithm almost halves the time spent by the algorithm in [14].…”

“…Son T et al [20] proposed a surface signature descriptor to recompose the fragments. It performs the initial matching the fragments according to the spin images, and then conducts the mismatch elimination according to RANSAC and the distance from the point to the line.…”

Pairwise Matching for 3D Fragment Reassembly Based on Boundary Curves and Concave-Convex Patches

“…In this work, we focus mainly on the pairwise matching of fragments. Based on our pairwise matching, we use the simple and greedy method in [5] to globally reassemble fragments. It is used to evaluate the performance of our method.…”

“…Recently, studies [3]- [13] matched fragments based on features of the fracture surface, which used feature patches [4], [5], feature curves [3], [6], [8], and feature points [9]. For example, Huang et al [4] clustered points to generate…”