CurveNet:Curvature-Based Multitask Learning Deep Networks for 3D Object Recognition

作     者：A.A.M.Muzahid Wanggen Wan Ferdous Sohel Lianyao Wu Li Hou A.A.M.Muzahid;Wanggen Wan;Ferdous Sohel;Lianyao Wu;Li Hou

作者机构：School of Communications and Information EngineeringInstitute of Smart CityShanghai UniversityShanghai 200444China Discipline of Information TechnologyMurdoch UniversityMurdoch WA 6150Australia School of Information EngineeringHuangshan UniversityHuangshan 245041China 

出 版 物：《IEEE/CAA Journal of Automatica Sinica》 (自动化学报（英文版）)

年 卷 期：2021年第8卷第6期

页      面：1177-1187页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 080203[工学-机械设计及理论] 0835[工学-软件工程] 0802[工学-机械工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：This paper was partially supported by a project of the Shanghai Science and Technology Committee(18510760300) Anhui Natural Science Foundation(1908085MF178) Anhui Excellent Young Talents Support Program Project(gxyqZD2019069) 

主　　题：3D shape analysis convolutional neural network DNNs object classification volumetric CNN 

摘      要：In computer vision fields,3D object recognition is one of the most important tasks for many real-world ***-dimensional convolutional neural networks(CNNs)have demonstrated their advantages in 3D object *** this paper,we propose to use the principal curvature directions of 3D objects(using a CAD model)to represent the geometric features as inputs for the 3D *** framework,namely CurveNet,learns perceptually relevant salient features and predicts object class *** directions incorporate complex surface information of a 3D object,which helps our framework to produce more precise and discriminative features for object *** learning is inspired by sharing features between two related tasks,where we consider pose classification as an auxiliary task to enable our CurveNet to better generalize object label *** results show that our proposed framework using curvature vectors performs better than voxels as an input for 3D object *** further improved the performance of CurveNet by combining two networks with both curvature direction and voxels of a 3D object as the inputs.A Cross-Stitch module was adopted to learn effective shared features across multiple *** evaluated our methods using three publicly available datasets and achieved competitive performance in the 3D object recognition task.