A performance evaluation of point pair features

被引:25
作者
Kiforenko, Lilita [1 ]
Drost, Bertram [2 ]
Tombari, Federico [3 ]
Kruger, Norbert [1 ]
Buch, Anders Glent [1 ]
机构
[1] Univ Southern Denmark, Maersk Mc Kinney Moller Inst, Odense, Denmark
[2] MVTec Software GmbH, Munich, Germany
[3] Tech Univ Munich, Garching, Germany
关键词
PPF; Point pair features; Object detection; Object recognition; Pose estimation; Feature description; 3D OBJECT RECOGNITION; POSE ESTIMATION; IMAGES;
D O I
10.1016/j.cviu.2017.09.004
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
More than a decade ago, the point pair features (PPFs) were introduced, showing a great potential for 3D object detection and pose estimation under very different conditions. Many modifications have been made to the original PPF, in each case showing varying degrees of improvement for specific datasets. However, to the best of our knowledge, no comprehensive evaluation of these features has been made. In this work, we evaluate PPFs on a large set of 3D scenes. We not only compare PPFs to local point cloud descriptors, but also investigate the internal variations of PPFs (different types of relations between two points). Our comparison is made on 7 publicly available datasets, showing variations on a number of parameters, e.g. acquisition technique, the number of objects/scenes and the amount of occlusion and clutter. We evaluate feature performance both at a point-wise object-scene correspondence level and for overall object detection and pose estimation in a RANSAC pipeline. Additionally, we also present object detection and pose estimation results for the original, voting based, PPF algorithm. Our results show that in general PPF is the top performer, however, there are datasets, which have low resolution data, where local histogram features show a higher performance than PPFs. We also found that PPFs compared to most local histogram features degrade faster under disturbances such as occlusion and clutter, however, PPFs still remain more descriptive on an absolute scale. The main contribution of this paper is a detailed analysis of PPFs, which highlights under which conditions PPFs perform particularly well as well as its main weaknesses.
引用
收藏
页码:66 / 80
页数:15
相关论文
共 34 条
[1]  
[Anonymous], 2009, IEEE INT C ROB AUT
[2]  
[Anonymous], 2013, PROC ASIAN C COMPUT
[3]  
[Anonymous], P 10 INT C COMP VIS
[4]  
Bayramoglu Neslihan, 2010, Proceedings of the 2010 20th International Conference on Pattern Recognition (ICPR 2010), P352, DOI 10.1109/ICPR.2010.95
[5]   A METHOD FOR REGISTRATION OF 3-D SHAPES [J].
BESL, PJ ;
MCKAY, ND .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1992, 14 (02) :239-256
[6]   Point Pair Features Based Object Detection and Pose Estimation Revisited [J].
Birdal, Tolga ;
Ilic, Slobodan .
2015 INTERNATIONAL CONFERENCE ON 3D VISION, 2015, :527-535
[7]   Local shape feature fusion for improved matching, pose estimation and 3D object recognition [J].
Buch, Anders G. ;
Petersen, Henrik G. ;
Kruger, Norbert .
SPRINGERPLUS, 2016, 5
[8]   RGB-D object pose estimation in unstructured environments [J].
Choi, Changhyun ;
Christensen, Henrik I. .
ROBOTICS AND AUTONOMOUS SYSTEMS, 2016, 75 :595-613
[9]  
Choi C, 2013, IEEE INT C INT ROBOT, P1568, DOI 10.1109/IROS.2013.6696558
[10]  
Choi C, 2012, IEEE INT C INT ROBOT, P3342, DOI 10.1109/IROS.2012.6386067