Matches in SemOpenAlex for { <https://semopenalex.org/work/W2783241901> ?p ?o ?g. }
- W2783241901 abstract "Deep learning methods have become the default tool for image classification. However, application of deep learning to surface shape classification is burdened by the limitations of existing methods, in particular, by lack of invariance to geometric transformations of input data.This thesis proposes two novel frameworks for learning a multi-layer representation of surface shape features, namely the view-based and the surface-based compositional hierarchical frameworks. The proposed representation is a hierarchical vocabulary of shape features, termed parts. Parts of the first layer are pre-defined, while parts of the subsequent layers, describing spatial relations of subparts, are learned. The view-based framework describes spatial relations between subparts using a camera-based reference frame. The key stage of the learning algorithm is part selection which forms the vocabulary based on multi-objective optimization, considering different importance measures of parts. Our experiments show that this framework enables efficient category recognition on a large-scale dataset.The surface-based framework exploits part-based intrinsic reference frames, which are computed for lower layers parts and inherited by parts of the subsequent layers. During learning spatial relations between subparts are described in these reference frames. During inference, a part is detected in input data when its subparts are detected at certain positions and orientations in each other’s reference frames. Since rigid body transformations don’t change positions and orientations of parts in intrinsic reference frames, this approach enables efficient recognition from unseen poses. Experiments show that this framework exhibits a large discriminative power and greater robustness to rigid body transformations than advanced CNN-based methods." @default.
- W2783241901 created "2018-01-26" @default.
- W2783241901 creator A5035310680 @default.
- W2783241901 date "2017-12-01" @default.
- W2783241901 modified "2023-09-23" @default.
- W2783241901 title "3D compositional hierarchies for object categorization" @default.
- W2783241901 cites W1487718662 @default.
- W2783241901 cites W1488482743 @default.
- W2783241901 cites W1488569956 @default.
- W2783241901 cites W1490682695 @default.
- W2783241901 cites W1507347479 @default.
- W2783241901 cites W1540596182 @default.
- W2783241901 cites W1561952261 @default.
- W2783241901 cites W1566474797 @default.
- W2783241901 cites W1566944040 @default.
- W2783241901 cites W1593727536 @default.
- W2783241901 cites W1596318652 @default.
- W2783241901 cites W1596547119 @default.
- W2783241901 cites W1619816297 @default.
- W2783241901 cites W1629010235 @default.
- W2783241901 cites W1686810756 @default.
- W2783241901 cites W1699734612 @default.
- W2783241901 cites W1798731418 @default.
- W2783241901 cites W1813659000 @default.
- W2783241901 cites W1821057786 @default.
- W2783241901 cites W1849277567 @default.
- W2783241901 cites W1885760432 @default.
- W2783241901 cites W1931582241 @default.
- W2783241901 cites W1951806617 @default.
- W2783241901 cites W1955462214 @default.
- W2783241901 cites W1963610420 @default.
- W2783241901 cites W1965080940 @default.
- W2783241901 cites W1969081995 @default.
- W2783241901 cites W1984475609 @default.
- W2783241901 cites W1986482242 @default.
- W2783241901 cites W1989625560 @default.
- W2783241901 cites W1991752086 @default.
- W2783241901 cites W1994205671 @default.
- W2783241901 cites W2001353114 @default.
- W2783241901 cites W2007200979 @default.
- W2783241901 cites W2007206727 @default.
- W2783241901 cites W2007876834 @default.
- W2783241901 cites W2011798472 @default.
- W2783241901 cites W2016215417 @default.
- W2783241901 cites W2016423476 @default.
- W2783241901 cites W2020682184 @default.
- W2783241901 cites W2023056405 @default.
- W2783241901 cites W2032533296 @default.
- W2783241901 cites W2051224630 @default.
- W2783241901 cites W2051730763 @default.
- W2783241901 cites W2052916917 @default.
- W2783241901 cites W2054113582 @default.
- W2783241901 cites W2054329862 @default.
- W2783241901 cites W2054658115 @default.
- W2783241901 cites W2057960707 @default.
- W2783241901 cites W2058605740 @default.
- W2783241901 cites W2059630294 @default.
- W2783241901 cites W2059917035 @default.
- W2783241901 cites W2065585006 @default.
- W2783241901 cites W2068470708 @default.
- W2783241901 cites W2070216603 @default.
- W2783241901 cites W2074142320 @default.
- W2783241901 cites W2074781138 @default.
- W2783241901 cites W2079363745 @default.
- W2783241901 cites W2085261163 @default.
- W2783241901 cites W2085411191 @default.
- W2783241901 cites W2087353192 @default.
- W2783241901 cites W2091791686 @default.
- W2783241901 cites W2091845343 @default.
- W2783241901 cites W2094262711 @default.
- W2783241901 cites W2096077837 @default.
- W2783241901 cites W2098580305 @default.
- W2783241901 cites W2098764590 @default.
- W2783241901 cites W2099606917 @default.
- W2783241901 cites W2099907898 @default.
- W2783241901 cites W2099972257 @default.
- W2783241901 cites W2100495367 @default.
- W2783241901 cites W2100510758 @default.
- W2783241901 cites W21006490 @default.
- W2783241901 cites W2101902009 @default.
- W2783241901 cites W2102605133 @default.
- W2783241901 cites W2106066189 @default.
- W2783241901 cites W2108598243 @default.
- W2783241901 cites W2108704430 @default.
- W2783241901 cites W2109992539 @default.
- W2783241901 cites W2110099550 @default.
- W2783241901 cites W2111457156 @default.
- W2783241901 cites W2114138863 @default.
- W2783241901 cites W2114846024 @default.
- W2783241901 cites W2115592720 @default.
- W2783241901 cites W2115782586 @default.
- W2783241901 cites W2116360511 @default.
- W2783241901 cites W2116752171 @default.
- W2783241901 cites W2117183049 @default.
- W2783241901 cites W2117539524 @default.
- W2783241901 cites W2117731089 @default.
- W2783241901 cites W2118010438 @default.
- W2783241901 cites W2119617729 @default.
- W2783241901 cites W2120419212 @default.
- W2783241901 cites W2120480077 @default.