Seite - 18 - in Joint Austrian Computer Vision and Robotics Workshop 2020

Figure8.Visualizationof the results with the objectbeingplacedatdifferent anglesanddistances around thecamera lies shown in the data can be explained through poorly chosen templates. The problems faced could be solved in future work by using the recommenda- tionsgiven in thiswork. References [1] H. Abdi and L. J. Williams. Principal compo- nent analysis. WIREs Computational Statistics, 2(4):433–459,2010. [2] D. I. Barnea and H. F. Silverman. A Class of Al- gorithms for Fast Digital Image Registration. IEEE Transactions on Computers, C-21(2):179–186, Feb. 1972. [3] P.Besl andN. D.McKay. Amethod for registration of 3-D shapes. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 14(2):239–256, Feb. 1992. [4] L. Cole, D. Austin, and L. Cole. Visual Object Recognition using Template Matching. In Proceed- ings of Australian Conference on Robotics and Au- tomation, 2004. [5] R. M. Dufour, E. L. Miller, and N. P. Galatsanos. Template matching based object recognition with unknown geometric parameters. IEEE Transactions on ImageProcessing, 11(12):1385–1396,2002. [6] S. Korman, D. Reichman, G. Tsur, and S. Avidan. Fast-Match: Fast Affine Template Matching. Inter- national Journal of Computer Vision, 121(1):111– 125, Jan.2017. [7] K.Lai,L.Bo,X.Ren, andD.Fox. Alarge-scalehi- erarchical multi-view rgb-d object dataset. In 2011 IEEE international conference on robotics and au- tomation, pages 1817–1824. IEEE,2011. [8] Y. Lu and D. Song. Robustness to lighting varia- tions: AnRGB-Dindoorvisualodometryusing line segments. In 2015 IEEE/RSJ International Con- ference on Intelligent Robots and Systems (IROS), pages688–694. IEEE,2015. [9] G. Marcus. Deep learning: A critical appraisal. arXivpreprintarXiv:1801.00631, 2018. [10] V. Nabat, M. de la O RODRIGUEZ, O. Company, S.Krut,andF.Pierrot.Par4: veryhighspeedparallel robotforpick-and-place. In2005IEEE/RSJInterna- tional conference on intelligent robots and systems, pages 553–558. IEEE, 2005. [11] R.Opromolla,G.Fasano,G.Rufino,andM.Grassi. A model-based 3d template matching technique for pose acquisition of an uncooperative space object. Sensors, 15(3):6360–6382,2015. [12] F. Pomerleau, F. Colas, R. Siegwart, and S. Magne- nat. ComparingICPvariantsonreal-worlddatasets. AutonomousRobots, 34(3):133–148,Apr.2013. [13] J. N. Rauer. Semi-Automatic Generation of Train- ing Data for Neural Networks for 6d Pose Esti- mation and Robotic Grasping. Master’s thesis, FachhochschuleTechnikumWien,Ho¨chsta¨dtplatz5, 1200Wien,2019. [14] S. Rusinkiewicz and M. Levoy. Efficient variants of the ICP algorithm. In Proceedings Third Interna- tionalConferenceon3-DDigital ImagingandMod- eling, pages 145–152,May2001. ISSN:null. [15] C. Szegedy, A. Toshev, and D. Erhan. Deep neural networksforobjectdetection. InAdvances inneural information processing systems, pages 2553–2561, 2013. [16] B.S.TjanandG.E.Legge. Theviewpointcomplex- ity of an object-recognition task. Vision Research, 38(15):2335–2350,Aug.1998. [17] F. Ullah and S. Kaneko. Using orientation codes for rotation-invariant templatematching. Patternrecog- nition, 37(2):201–209,2004. [18] P.WohlhartandV.Lepetit. LearningDescriptors for Object Recognition and 3d Pose Estimation. Pro- ceedings IEEE Conference on Computer Vision and PatternRecognition (CVPR), Feb.2015. [19] Y. Xiang, T. Schmidt, V. Narayanan, and D. Fox. PoseCNN: A Convolutional Neural Network for 6d Object Pose Estimation in Cluttered Scenes. arXiv:1711.00199 [cs], May 2018. arXiv: 1711.00199. 18