Klasifikace na množinách bodů v 3D
| Název práce v češtině: | Klasifikace na množinách bodů v 3D |
|---|---|
| Název v anglickém jazyce: | Classification on point sets in 3D |
| Klíčová slova: | hluboké učení, klasifikace, neuronové sítě |
| Klíčová slova anglicky: | deep learning, classification, neural networks |
| Akademický rok vypsání: | 2017/2018 |
| Typ práce: | diplomová práce |
| Jazyk práce: | angličtina |
| Ústav: | Katedra softwaru a výuky informatiky (32-KSVI) |
| Vedoucí / školitel: | RNDr. František Mráz, CSc. |
| Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
| Datum přihlášení: | 05.02.2018 |
| Datum zadání: | 08.02.2018 |
| Datum potvrzení stud. oddělením: | 20.07.2018 |
| Datum a čas obhajoby: | 13.09.2018 09:00 |
| Datum odevzdání elektronické podoby: | 17.07.2018 |
| Datum odevzdání tištěné podoby: | 20.07.2018 |
| Datum proběhlé obhajoby: | 13.09.2018 |
| Oponenti: | doc. RNDr. Elena Šikudová, Ph.D. |
| Zásady pro vypracování |
| Neural networks were successfully applied to object classification on sets of points in 3D. Qi, Su, Mo a Guibas [3,4] designed an architecture of deep network called PointNet, which they apply to classification and segmentation of sets of points in 3D. This architecture is based on computing new attributes using symmetric functions of point coordinates. These new attributes are then used in the proper classification. The computations of the new attributes for a point set did not use any information about local structure of neighbourhood of points like, e.g., density or direction to their closest neighbours. The goal of the thesis is to propose an extension of the input for object classification by adding such new attributes describing neighbourhood of each input point in 3D and to compare the results of suitable models of deep networks on selected classification problems with the new attributes and without them. |
| Seznam odborné literatury |
| [1] Hofer, C., Kwitt, R., Niethammer, M., & Uhl, A.: Deep Learning with Topological Signatures. In Advances in Neural Information Processing Systems. 2017, pp. 1633-1643.
[2] Klokov, R., & Lempitsky, V.: Escape from cells: Deep kd-networks for the recognition of 3d point cloud models. In 2017 IEEE International Conference on Computer Vision (ICCV), IEEE, 2017, pp. 863-872. [3] Qi, C. R., Su, H., Mo, K., & Guibas, L. J.: Pointnet: Deep learning on point sets for 3d classification and segmentation. Proc. Computer Vision and Pattern Recognition (CVPR), IEEE, 2017, 1(2), 4. [4] Qi, C. R., Yi, L., Su, H., & Guibas, L. J.: Pointnet++: Deep hierarchical feature learning on point sets in a metric space. In Advances in Neural Information Processing Systems, 2017, pp. 5105-5114. |