Please use this identifier to cite or link to this item: http://hdl.handle.net/2440/111360
Citations
Scopus Web of Science® Altmetric
?
?
Type: Conference paper
Title: Scaling CNNs for high resolution volumetric reconstruction from a single image
Author: Johnston, A.
Garg, R.
Carneiro, G.
Reid, I.
van den Hengel, A.
Citation: Proceedings of the IEEE International Conference on Computer Vision Workshop (ICCVW 2017), 2017 / vol.2018-January, pp.930-939
Publisher: IEEE
Publisher Place: Piscataway, NJ
Issue Date: 2017
Series/Report no.: IEEE International Conference on Computer Vision Workshops
ISBN: 9781538610350
ISSN: 2473-9936
Conference Name: IEEE International Conference on Computer Vision Workshop (ICCVW 2017) (22 Oct 2017 - 29 Oct 2017 : Venice, ITALY)
Statement of
Responsibility: 
Adrian Johnston, Ravi Garg, Gustavo Carneiro, Ian Reid, Anton van den Hengel
Abstract: One of the long-standing tasks in computer vision is to use a single 2-D view of an object in order to produce its 3-D shape. Recovering the lost dimension in this process has been the goal of classic shape-from-X methods, but often the assumptions made in those works are quite limiting to be useful for general 3-D objects. This problem has been recently addressed with deep learning methods containing a 2-D (convolution) encoder followed by a 3-D (deconvolution) decoder. These methods have been reasonably successful, but memory and run time constraints impose a strong limitation in terms of the resolution of the reconstructed 3-D shapes. In particular, state-of-the-art methods are able to reconstruct 3-D shapes represented by volumes of at most 323 voxels using state-of-the-art desktop computers. In this work, we present a scalable 2-D single view to 3-D volume reconstruction deep learning method, where the 3-D (deconvolution) decoder is replaced by a simple inverse discrete cosine transform (IDCT) decoder. Our simpler architecture has an order of magnitude faster inference when reconstructing 3-D volumes compared to the convolution-deconvolutional model, an exponentially smaller memory complexity while training and testing, and a sub-linear run-time training complexity with respect to the output volume size. We show on benchmark datasets that our method can produce high-resolution reconstructions with state of the art accuracy.
Rights: © 2017 IEEE
RMID: 0030083745
DOI: 10.1109/ICCVW.2017.114
Grant ID: http://purl.org/au-research/grants/arc/CE140100016
http://purl.org/au-research/grants/arc/FL130100102
Published version: http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=8234943
Appears in Collections:Computer Science publications

Files in This Item:
There are no files associated with this item.


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.