TY - JOUR
T1 - Rotational contour signatures for both real-valued and binary feature representations of 3D local shape
AU - Yang, Jiaqi
AU - Zhang, Qian
AU - Xian, Ke
AU - Xiao, Yang
AU - Cao, Zhiguo
N1 - Publisher Copyright:
© 2017 Elsevier Inc.
PY - 2017/7
Y1 - 2017/7
N2 - This paper presents a rotational contour signatures (RCS) method for both real-valued and binary descriptions of 3D local shape. RCS comprises several signatures that characterize the 2D contour information derived from 3D-to-2D projection of the local point cloud. The inspiration of our encoding technique comes from that when viewing towards an object, its contour is an effective and robust cue for representing its shape. In order to achieve a comprehensive geometry encoding, the local surface is continually rotated in a predefined local reference frame (LRF) so that multi-view information is obtained. A peculiar trait of our RCS method is its seamless extension to binary representations to accelerate feature matching and reduce storage consumption. Specifically, we resort to three techniques, i.e., thresholding, quantization and geometrical binary encoding, to generate RCS binary strings. In contrast to 2D image area, there are quite rare 3D binary descriptors yet in 3D computer vision. We deploy experiments on three standard datasets including shape retrieval, 3D object recognition and 2.5D point cloud view matching scenarios with a rigorous comparison with six state-of-the-art descriptors. The comparative outcomes confirm numerous merits of our RCS method, e.g., highly discriminative, compact, computational efficient and robust to many nuisances including noise, mesh resolution variation, clutter and occlusion. We also show the versatility of RCS in matching of both LiDAR and Kinect point clouds.
AB - This paper presents a rotational contour signatures (RCS) method for both real-valued and binary descriptions of 3D local shape. RCS comprises several signatures that characterize the 2D contour information derived from 3D-to-2D projection of the local point cloud. The inspiration of our encoding technique comes from that when viewing towards an object, its contour is an effective and robust cue for representing its shape. In order to achieve a comprehensive geometry encoding, the local surface is continually rotated in a predefined local reference frame (LRF) so that multi-view information is obtained. A peculiar trait of our RCS method is its seamless extension to binary representations to accelerate feature matching and reduce storage consumption. Specifically, we resort to three techniques, i.e., thresholding, quantization and geometrical binary encoding, to generate RCS binary strings. In contrast to 2D image area, there are quite rare 3D binary descriptors yet in 3D computer vision. We deploy experiments on three standard datasets including shape retrieval, 3D object recognition and 2.5D point cloud view matching scenarios with a rigorous comparison with six state-of-the-art descriptors. The comparative outcomes confirm numerous merits of our RCS method, e.g., highly discriminative, compact, computational efficient and robust to many nuisances including noise, mesh resolution variation, clutter and occlusion. We also show the versatility of RCS in matching of both LiDAR and Kinect point clouds.
KW - Binary representation
KW - Contour signature
KW - Feature matching
KW - Local shape descriptor
KW - Rotation
UR - http://www.scopus.com/inward/record.url?scp=85013498868&partnerID=8YFLogxK
U2 - 10.1016/j.cviu.2017.02.004
DO - 10.1016/j.cviu.2017.02.004
M3 - 文章
AN - SCOPUS:85013498868
SN - 1077-3142
VL - 160
SP - 133
EP - 147
JO - Computer Vision and Image Understanding
JF - Computer Vision and Image Understanding
ER -