File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)

Conference Paper: Data-driven light field depth estimation using deep convolution neural network

TitleData-driven light field depth estimation using deep convolution neural network
Authors
Issue Date2016
PublisherIEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000500
Citation
The 2016 International Joint Conference on Neural Networks (IJCNN 2016), Vancouver, Canada, 24-29 July 2016. In Conference Proceedings, 2016, p. 367-374 How to Cite?
AbstractThis paper presents a data-driven approach to estimate the object depths from light field data using Convolutional Neural Networks (CNN). By exploring the relationship between the epipolar-plane images (EPI) and the corresponding depth map, we propose an enhanced EPI feature that encodes the depth information of each physical point in the light field and obtains the disparity map of the whole scene in a supervised manner. This work covers two major contributions, namely the extraction of the enhanced EPI features and the light field depth estimation with CNN. The proposed features augment the depth information of the corresponding points in the light field, and then our CNN architecture differentiates them into different depth layers. Forward propagation step of the CNN model allows rapid recognition of the disparity map of the test light field data. In the experiments, we apply our method on the HCI (Heidelberg Col-laboratory for Image Processing) benchmark dataset and demonstrate that it is significantly faster than the state-of-the-art light field depth estimation approaches while achieving satisfactory performance.
DescriptionIEEE WCCI 2016 will host three conferences: The 2016 International Joint Conference on Neural Networks (IJCNN 2016), the 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2016), and the 2016 IEEE Congress on Evolutionary Computation (IEEE CEC 2016)
Persistent Identifierhttp://hdl.handle.net/10722/234985
ISSN

 

DC FieldValueLanguage
dc.contributor.authorSun, X-
dc.contributor.authorXu, Z-
dc.contributor.authorMeng, N-
dc.contributor.authorLam, EYM-
dc.contributor.authorSo, HKH-
dc.date.accessioned2016-10-14T13:50:32Z-
dc.date.available2016-10-14T13:50:32Z-
dc.date.issued2016-
dc.identifier.citationThe 2016 International Joint Conference on Neural Networks (IJCNN 2016), Vancouver, Canada, 24-29 July 2016. In Conference Proceedings, 2016, p. 367-374-
dc.identifier.issn2161-4407-
dc.identifier.urihttp://hdl.handle.net/10722/234985-
dc.descriptionIEEE WCCI 2016 will host three conferences: The 2016 International Joint Conference on Neural Networks (IJCNN 2016), the 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2016), and the 2016 IEEE Congress on Evolutionary Computation (IEEE CEC 2016)-
dc.description.abstractThis paper presents a data-driven approach to estimate the object depths from light field data using Convolutional Neural Networks (CNN). By exploring the relationship between the epipolar-plane images (EPI) and the corresponding depth map, we propose an enhanced EPI feature that encodes the depth information of each physical point in the light field and obtains the disparity map of the whole scene in a supervised manner. This work covers two major contributions, namely the extraction of the enhanced EPI features and the light field depth estimation with CNN. The proposed features augment the depth information of the corresponding points in the light field, and then our CNN architecture differentiates them into different depth layers. Forward propagation step of the CNN model allows rapid recognition of the disparity map of the test light field data. In the experiments, we apply our method on the HCI (Heidelberg Col-laboratory for Image Processing) benchmark dataset and demonstrate that it is significantly faster than the state-of-the-art light field depth estimation approaches while achieving satisfactory performance.-
dc.languageeng-
dc.publisherIEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000500-
dc.relation.ispartofInternational Joint Conference on Neural Networks (IJCNN) Proceedings-
dc.rightsInternational Joint Conference on Neural Networks (IJCNN) Proceedings. Copyright © IEEE.-
dc.rights©2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.-
dc.titleData-driven light field depth estimation using deep convolution neural network-
dc.typeConference_Paper-
dc.identifier.emailLam, EYM: elam@eee.hku.hk-
dc.identifier.emailSo, HKH: hso@eee.hku.hk-
dc.identifier.authorityLam, EYM=rp00131-
dc.identifier.authoritySo, HKH=rp00169-
dc.identifier.doi10.1109/IJCNN.2016.7727222-
dc.identifier.hkuros268707-
dc.identifier.spage367-
dc.identifier.epage374-
dc.publisher.placeUnited States-
dc.customcontrol.immutablesml 161202-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats