Classification of occluded 2D objects using deep learning of 3D shape surfaces

被引：0

作者：

Tzitzilonis, Vasileios ^{[1
]}

Kappatos, Vassilios ^{[2
]}

Dermatas, Evangelos ^{[1
]}

Apostolopoulos, George ^{[1
]}

机构：

[1] Univ Patras, Dept Elect Engn & Comp Technol, Patras 26500, Greece

[2] Univ Southern Denmark, Dept Technol & Innovat ITI, DK-5230 Odense, Denmark

来源：

10TH HELLENIC CONFERENCE ON ARTIFICIAL INTELLIGENCE (SETN 2018) | 2018年

关键词：

Convolutional Neural Networks; Classification; 3D-Models; Deep Learning; Computer Vision; Machine learning;

D O I：

10.1145/3200947.3201061

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents a novel deep learning method for partially occluded 2D object classification. A 2D Convolutional Neural Network (CNN) was trained with partial and whole images of the 3D models obtained from different camera views. The efficiency of the proposed method in classifying partial objects in 40 categories is more than 80% in most objects and exceeds 95% in some of them.

引用

页数：2

共 7 条

[1]

[Anonymous], 2015, PROC CVPR IEEE, DOI 10.1109/CVPR.2015.7298801

[2]

[Anonymous], COMPUTER VISION PATT

[3]

[Anonymous], 2015, 3 INT C LEARNING REP

[4]

Chollet F., 2015, about us

[5]

Ioffe Sergey, 2015, P MACHINE LEARNING R, V37, P448, DOI [DOI 10.48550/ARXIV.1502.03167, DOI 10.5555/3015118.3045167]

[6] 3-D Model Search and Retrieval From Range Images Using Salient Features [J].

Stavropoulos, Georgios ;

Moschonas, Panagiotis ;

Moustakas, Konstantinos ;

Tzovaras, Dimitrios ;

Strintzis, Michael Gerassimos .

IEEE TRANSACTIONS ON MULTIMEDIA, 2010, 12 (07) :692-704

[7] Multi-view Convolutional Neural Networks for 3D Shape Recognition [J].

Su, Hang ;

Maji, Subhransu ;

Kalogerakis, Evangelos ;

Learned-Miller, Erik .

2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2015, :945-953

← 1 →