Title | ||
---|---|---|
Hough-space-based hypothesis generation and hypothesis verification for 3D object recognition and 6D pose estimation. |
Abstract | ||
---|---|---|
•We use an only-2D Hough space to generate a Hough accumulator for a 6DoF hypothesis generation.•Our “double RANSAC” eliminates false matching correspondences, based on feature distance of correspondences and point position’s geometric consistency of correspondences.•We design a self-adapted approach to decrease the influence from fixed parameter setting in hypothesis generation.•We introduce fitness score and matching rate for our double-ICP-based hypothesis verification which can recognize objects under high occlusion or high false matching correspondences.•We introduce “FP vs Recognition curve” to check the performance of recognition in terms of different False Positive rate of input correspondences. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1016/j.cag.2018.01.011 | Computers & Graphics |
Keywords | Field | DocType |
Local surface patch,Hypothesis Verification,Local feature descriptor,Hypothesis Generation,3D object recognition | Computer vision,Computer science,Hough space,Clutter,Pose,Artificial intelligence,Hypothesis verification,Cognitive neuroscience of visual object recognition | Journal |
Volume | ISSN | Citations |
72 | 0097-8493 | 2 |
PageRank | References | Authors |
0.37 | 29 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wei Zhou | 1 | 15 | 2.92 |
Caiwen Ma | 2 | 2 | 0.37 |
Arjan Kuijper | 3 | 1063 | 133.22 |