Graduate School and Research Center in Digital Sciences

A new framework for optimal facial landmark localization on light-field images

Galdi, Chiara; Younes, Lara; Guillemot, Christine; Dugelay, Jean-Luc

VCIP 2018, IEEE International Conference on Visual Communications and Image Processing, December 9-12, 2018, Taichung, Taiwan

The paper explores how light fields captured by plenoptic cameras can increase the performance of face landmark detection. The idea is to exploit light fields geometrical constraints to correct the position of points detected by classical face landmark detectors. These geometric constraints are used to enforce landmark points angular coherency across the different views of the light field, and by doing so to correct the positions of the landmarks on all views. The corrected landmark points are compared with ground-truth manual annotations of a set of 400 images corresponding to the central views of 400 light fields of faces with different pose and expression.

Document Doi Bibtex

Title:A new framework for optimal facial landmark localization on light-field images
Keywords:light fields, face landmark detection, structure tensor, LFFD, EPI
Type:Conference
Language:English
City:Taichung
Country:TAIWAN, PROVINCE OF CHINA
Date:
Department:Digital Security
Eurecom ref:5736
Copyright: © 2018 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Bibtex: @inproceedings{EURECOM+5736, doi = {http://dx.doi.org/10.1109/VCIP.2018.8698735}, year = {2018}, title = {{A} new framework for optimal facial landmark localization on light-field images}, author = {{G}aldi, {C}hiara and {Y}ounes, {L}ara and {G}uillemot, {C}hristine and {D}ugelay, {J}ean-{L}uc}, booktitle = {{VCIP} 2018, {IEEE} {I}nternational {C}onference on {V}isual {C}ommunications and {I}mage {P}rocessing, {D}ecember 9-12, 2018, {T}aichung, {T}aiwan}, address = {{T}aichung, {TAIWAN}, {PROVINCE} {OF} {CHINA}}, month = {12}, url = {http://www.eurecom.fr/publication/5736} }
See also: