Globfit: Consistently fitting primitives by discovering global relations
Li, Yangyan; Wu, Xiaokun; Chrysathou, Yiorgos; Sharf, Andrei; Cohen-Or, Daniel; Mitra, Niloy J.
2011
会议名称ACM SIGGRAPH 2011
会议地点加拿大温哥华
英文摘要Given a noisy and incomplete point set, we introduce a method that simultaneously recovers a set of locally fitted primitives along with their global mutual relations. We operate under the assumption that the data corresponds to a man-made engineering object consisting of basic primitives, possibly repeated and globallyaligned under common relations. We introduce an algorithm to directly couple the local and global aspects of the problem. The local fit of the model is determined byhow well the inferred model agrees to the observed data, while the global relations are iteratively learned and enforced through a constrained optimization. Starting with a set of initial RANSAC based locally fitted primitives, relations across the primitives such as orientation, placement, and equality are progressively learned and conformed to. In each stage, a set of feasible relations are extracted among the candidate relations, and then aligned to, while best fitting to the input data. Theglobal coupling corrects the primitives obtained in the local RANSAC stage, and brings them to precise global alignment. We test the robustness of our algorithm on a range of synthesized and scanned data, with varying amounts of noise, outliers, and non-uniform sampling, and validate the results against ground truth, where available. © 2011 ACM.(29 refs)
收录类别EI
语种英语
内容类型会议论文
源URL[http://ir.siat.ac.cn:8080/handle/172644/3593]  
专题深圳先进技术研究院_数字所
作者单位2011
推荐引用方式
GB/T 7714
Li, Yangyan,Wu, Xiaokun,Chrysathou, Yiorgos,et al. Globfit: Consistently fitting primitives by discovering global relations[C]. 见:ACM SIGGRAPH 2011. 加拿大温哥华.
个性服务
查看访问统计
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。


©版权所有 ©2017 CSpace - Powered by CSpace