Performance of distance-based matching algorithms in 3D facial
identification

J 2016

Performance of distance-based matching algorithms in 3D facial identification

URBANOVÁ, Petra

Basic information

Original name

Performance of distance-based matching algorithms in 3D facial identification

Authors

URBANOVÁ, Petra (203 Czech Republic, guarantor, belonging to the institution)

Edition

Egyptian Journal of Forensic Sciences, Egypt, Elsevier, 2016, 2090-536X

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

30501 Forensic science

Country of publisher

Egypt

Confidentiality degree

není předmětem státního či obchodního tajemství

References:

URL

RIV identification code

RIV/00216224:14310/16:00090192

Organization unit

Faculty of Science

DOI

http://dx.doi.org/10.1016/j.ejfs.2016.04.004

Keywords in English

3D facial identification; 3D meshes; Landmarks; Performance rate

Abstract

V originále

Facial image identification is an area of forensic sciences, where an expert provides an opinion on whether or not two or more images depict the same individual. The primary concern for facial image identification is that it must be based on sound scientific principles. The recent extensive development in 3D recording technology, which is presumed to enhance performances of identification tasks, has made essential to question conditions, under which 3D images can yield accurate and reliable results. The present paper explores the effect of mesh resolution, adequacy of selected measures of dissimilarity and number of variables employed to encode identity-specific facial features on a dataset of 528 3D face models sampled from the Fidentis 3D Face Database (N=2100). In order to match 3D images two quantitative approaches were tested, the first based on closest point-to-point distances computed from registered surface models and the second grounded on Procrustes distances derived from discrete 3D facial points collected manually on textured 3D facial models. The results expressed in terms of rank-1 identification rates, ROC curves and likelihood ratios show that under optimized conditions the tested algorithms have the capacity to provide very accurate and reliable results. The performance of the tested algorithms is, however, highly dependent on mesh resolution and the number of variables employed in the task. The results also show that in addition to numerical measures of dissimilarity, various 3D visualization tools can be of assistance in the decision-making.

Links

MUNI/A/1379/2015, interní kód MU

Name: Rozvoj analytických nástrojů pro studium tělesných znaků člověka se zaměřením na analýzu dynamiky pohybu

Investor: Masaryk University, Category A

Citovat

URBANOVÁ, Petra. Performance of distance-based matching algorithms in 3D facial identification. Egyptian Journal of Forensic Sciences. Egypt: Elsevier, 2016, vol. 6, No 2, p. 135-151. ISSN 2090-536X. Available from: https://dx.doi.org/10.1016/j.ejfs.2016.04.004.

@article{1347723,
   author = {Urbanová, Petra},
   article_location = {Egypt},
   article_number = {2},
   doi = {http://dx.doi.org/10.1016/j.ejfs.2016.04.004},
   keywords = {3D facial identification; 3D meshes; Landmarks; Performance rate},
   language = {eng},
   issn = {2090-536X},
   journal = {Egyptian Journal of Forensic Sciences},
   title = {Performance of distance-based matching algorithms in 3D facial identification},
   url = {http://www.sciencedirect.com/science/article/pii/S2090536X16300089},
   volume = {6},
   year = {2016}
}

TY  - JOUR
ID  - 1347723
AU  - Urbanová, Petra
PY  - 2016
TI  - Performance of distance-based matching algorithms in 3D facial identification
JF  - Egyptian Journal of Forensic Sciences
VL  - 6
IS  - 2
SP  - 135-151
EP  - 135-151
PB  - Elsevier
SN  - 2090536X
KW  - 3D facial identification
KW  - 3D meshes
KW  - Landmarks
KW  - Performance rate
UR  - http://www.sciencedirect.com/science/article/pii/S2090536X16300089
N2  - Facial image identification is an area of forensic sciences, where an expert provides an opinion on whether or not two or more images depict the same individual. The primary concern for facial image identification is that it must be based on sound scientific principles. The recent extensive development in 3D recording technology, which is presumed to enhance performances of identification tasks, has made essential to question conditions, under which 3D images can yield accurate and reliable results. The present paper explores the effect of mesh resolution, adequacy of selected measures of dissimilarity and number of variables employed to encode identity-specific facial features on a dataset of 528 3D face models sampled from the Fidentis 3D Face Database (N=2100). In order to match 3D images two quantitative approaches were tested, the first based on closest point-to-point distances computed from registered surface models and the second grounded on Procrustes distances derived from discrete 3D facial points collected manually on textured 3D facial models. The results expressed in terms of rank-1 identification rates, ROC curves and likelihood ratios show that under optimized conditions the tested algorithms have the capacity to provide very accurate and reliable results. The performance of the tested algorithms is, however, highly dependent on mesh resolution and the number of variables employed in the task. The results also show that in addition to numerical measures of dissimilarity, various 3D visualization tools can be of assistance in the decision-making.
ER  -

URBANOVÁ, Petra. Performance of distance-based matching algorithms in 3D facial identification. \textit{Egyptian Journal of Forensic Sciences}. Egypt: Elsevier, 2016, vol.~6, No~2, p.~135-151. ISSN~2090-536X. Available from: https://dx.doi.org/10.1016/j.ejfs.2016.04.004.

Detailed Information on Publication Record