文献类型：期刊文献

英文题名：A Click-Based Interactive Segmentation Network for Point Clouds

作者：Sun, Wentao[1,2];Luo, Zhipeng[3];Chen, Yiping[4];Li, Huxiong[1,2];Marcato Jr, Jose[5];Gonealves, Wesley Nunes[5];Li, Jonathan[6,7]

机构：[1]Shaoxing Univ, Inst Artificial Intelligence, Shaoxing 312000, Zhejiang, Peoples R China;[2]Shaoxing Univ, Sch Mech & Elect Engn, Dept Comp Sci & Engn, Shaoxing 312000, Zhejiang, Peoples R China;[3]Minnan Normal Univ, Sch Comp Sci, Zhangzhou 363000, Fujian, Peoples R China;[4]Sun Yat Sen Univ, Sch Geospatial Engn & Sci, Zhuhai 519082, Guangdong, Peoples R China;[5]Univ Fed Mato Grosso do Sul, Fac Engn Architecture & Urbanism & Geog, BR-79070900 Campo Grande, Brazil;[6]Univ Waterloo, Dept Geog & Environm Management, Waterloo, ON N2L 3G1, Canada;[7]Univ Waterloo, Dept Syst Design Engn, Waterloo, ON N2L 3G1, Canada

年份：2023

卷号：61

起止页码：1

外文期刊名：IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING

收录：SCI-EXPANDED(收录号：WOS:001094836500013)、、EI(收录号：20234314968188)、Scopus(收录号：2-s2.0-85174847787)、WOS

基金：This work was supportedby the National Natural Science Foundation of China under Grant 42371343.

语种：英文

外文关键词：Deep learning; Deep learning (DL); instance segmentation; interactive segmentation; point clouds

外文摘要：Interactive segmentation plays an essential role in several tasks involving point clouds. However, existing methods suffer from low segmentation accuracy and cannot adjust the segmentation results according to the user's personal demands. This article presents a novel deep-learning (DL)-based interactive segmentation method, named click rough segmentation network (CRSNet), designed to handle point clouds. The method allows users to iteratively click to segment interesting objects. CRSNet consists of two key parts: a click rough segmentation (CRS) module and a feature extraction module. First, the CRS module transforms click operations into an appropriate representation to input into the feature extraction module. The CRS module takes raw point clouds and clicks operations as input and outputs 3-D Gaussian vectors and roughly segmented blocks, which adapt to different-sized and densely distributed objects in complex environments. Second, the feature extraction module, which uses a novel mix loss-based analysis algorithm, extracts deep features and obtains instance segmentation results. The module is highly compatible because its backbones can be replaced by different DL architectures. Experimental results on the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI), Apolloscape, Roadmarking, Scannet, and SemanticKITTI datasets show that our method outperforms state-of-the-art semantic segmentation methods with one click. Moreover, our method can generalize well to unseen objects and datasets.