Class-specific Object Pose Estimation and Reconstruction using 3D Part Geometry
File Type:
PDFItem Type:
Conference PaperDate:
2016Access:
openAccessCitation:
Kumar, A.C.S., Bodis-Szomuru, A., Bhandarkar, S. & Prasad, M., Class-specific Object Pose Estimation and Reconstruction using 3D Part Geometry, ECCV, Amsterdam, October 2016, 2016Download Item:
class-specific-object.pdf (PDF) 5.784Mb
Abstract:
We propose a novel approach for detecting and reconstructing class-specific objects from 2D images. Reconstruction and detection, despite major advances, are still wanting in performance. Hence, approaches that try to solve them jointly, so that one can be used to resolve the ambiguities of the other, especially while employing data-driven class-specific learning, are increasingly popular. In this paper, we learn a deformable, fine-grained, part-based model from real world, class-specific, image sequences, so that given a new image, we can simultaneously estimate the 3D shape, viewpoint and the subsequent 2D detection results. This is a step beyond existing approaches, which are usually limited to 3D CAD shapes, regression based pose estimation, template based deformation modelling etc. We employ Structure from Motion (SfM) and part based models in our learning process, and estimate a 3D deformable object instance and a projection matrix that explains the image information. We demonstrate our approach with high quality qualitative and quantitative results on our real world RealCar dataset, as well as the EPFL car dataset.
URI:
https://link.springer.com/chapter/10.1007/978-3-319-49409-8_23http://hdl.handle.net/2262/91812
Author's Homepage:
http://people.tcd.ie/prasadmOther Titles:
ECCVType of material:
Conference PaperURI:
https://link.springer.com/chapter/10.1007/978-3-319-49409-8_23http://hdl.handle.net/2262/91812
Collections:
Availability:
Full text availableSubject:
Radial basis function neural network, Convolutional neural networks, Appearance model, Angular error, Object instanceSubject (TCD):
Creative Technologies , RECONSTRUCTION , scene understandingThe following license files are associated with this item: