FACIAL SOFT BIOMETRICS - Library of Ph.D. Theses | EURASIP
FACIAL SOFT BIOMETRICS - Library of Ph.D. Theses | EURASIP FACIAL SOFT BIOMETRICS - Library of Ph.D. Theses | EURASIP
102 8. USER ACCEPTANCE STUDY RELATING TO SOFT BIOMETRICSApplying soft biometrics for person re-identification, with a focus on the frontal vs.side scenarioMotivated by realistic surveillance scenarios, the work addressed the problem of frontal-tosidefacial recognition, providing re–identification algorithms/classifiers that are specifically suitedfor this setting. Emphasis was placed on classifiers that belong in the class of soft biometric traits,specifically color–, texture– and intensity– based traits taken from patches of hair, skin and clothes.Towards providing insight, the work presented different identification experiments that adhere tothe frontal–to–side setting, as well as presented a preliminary analytical study that seeks to impartintuition on the role of the above traits in improving algorithmic reliability. Our analysis describedthe overall error probability, both as a function of collisions and of erroneous categorizations forgiven sizes of authentication groups. In the presence of a moderate reliability of the patches-basedmethod, the analysis suggests promising applications of this method in settings such as pruning ofsearches.Applying soft biometrics quantification and prediction of female facial aestheticsIn terms of female facial aesthetics, we presented a study on facial aesthetics in photographs,where we compared objective measures (namely photograph quality measures, facial beauty characteristicsand non permanent facial features), with human subjective perception. Our analysisrevealed a substantial correlation between different selected traits, and the corresponding MOSrelatedbeauty indices. Specifically we presented that non permanent features can influence highlythe MOS, and based on our analysis we conclude that facial aesthetics in images can indeedbe substantially modifiable. With other words parameters such as the presence of makeup andglasses, the image quality as well as different image post–processing methods can significantlyaffect the resulting MOS. Furthermore we constructed a linear MOS–based metric which wassuccessfully employed to quantify beauty-index variations due to aging and surgery. Our workapplies towards building a basis for designing new image-processing tools that further automateprediction of aesthetics in facial images. Towards this we provided a simulation of an automaticprediction tool based on state-of-art classification algorithms and the designed MOS–predictionmetric.The above approaches were accompanied by a more practically oriented part where we designedan automatic soft biometrics classification tool. Specifically we focused on eye, skin andhair color, as well as on the presence of beard, moustache and glasses.In terms of usability analysis, we presented a user study investigating the preference of a setof test participants on access methods, namely soft biometrics, face, PIN and fingerprint basedaccess methods. This preference was evaluated generally in terms of usability measures, suchas ease of use, intuitiveness and log-in-speed. Furthermore two scenarios were hereby assessed,specifically personal computer access and entrance of a security lab in a crowded environment.The surprising outcome is that although all users were strongly biased towards the PIN basedverification method, by daily use, the biometric based options were overall equally or even significantlybetter rated than the PIN based system. Users appreciated the comfort, easiness and speedof modern technology. Specifically they favored the soft biometrics system, due to the providedprivacy preservation and ease of use.
103Future WorkIt is becoming apparent that surveillance will increasingly affect our quality of life and security.Research in this area has been embraced by both academia and industry. For this reason,security related biometric systems will become larger and more dynamic. We see the area of softbiometrics having from now on a solid position in such systems. Towards this we will need betterunderstanding of the component parts of such SBSs, and a corresponding better understanding ofnovel trait classification algorithms, as well as novel ways of combining and analyzing such algorithms.Our aim will be to allow for more efficient SBSs, but also develop a rigorous understandingof the capabilities and limits of such systems.Our aim in the future will also be, in addition to developing novel algorithms for SBSs, to alsoidentify and develop new commercial applications that can benefit by the power of soft biometrics.
- Page 53 and 54: 51The following lemma describes the
- Page 55 and 56: 534.5.1 Typical behavior: average g
- Page 57 and 58: 55n = 50 subjects, out of which we
- Page 59 and 60: 5710.950.9pruning Gain r(vt)0.850.8
- Page 61 and 62: 59for one person, for trait t, t =
- Page 63 and 64: 61Chapter 5Frontal-to-side person r
- Page 65 and 66: 63Figure 5.1: Frontal / gallery and
- Page 67 and 68: 6510.90.80.7Skin colorHair colorShi
- Page 69 and 70: 6710.90.80.70.6Perr0.50.40.30.20.10
- Page 71 and 72: 69Chapter 6Soft biometrics for quan
- Page 73 and 74: 71raphy considerations include [BSS
- Page 75 and 76: 73Figure 6.3: Example image of the
- Page 77 and 78: 75A direct way to find a relationsh
- Page 79 and 80: 77- Pearson’s correlation coeffic
- Page 81 and 82: 79shown to have a high impact on ou
- Page 83 and 84: 81Chapter 7Practical implementation
- Page 85 and 86: 834) Eye glasses detection: Towards
- Page 87 and 88: 857.2 Eye color as a soft biometric
- Page 89 and 90: 87Table 7.5: GMM eye color results
- Page 91 and 92: 89and office lights, daylight, flas
- Page 93 and 94: 917.5 SummaryThis chapter presented
- Page 95 and 96: 93Chapter 8User acceptance study re
- Page 97 and 98: 95Table 8.1: User experience on acc
- Page 99 and 100: 97scared of their PIN being spying.
- Page 101 and 102: 99Table 8.2: Comparison of existing
- Page 103: 101ConclusionsThis dissertation exp
- Page 107 and 108: 105Appendix AAppendix for Section 3
- Page 109 and 110: 107- We are now left withN −F = 2
- Page 111 and 112: 109Appendix BAppendix to Section 4B
- Page 113 and 114: 111Blue Green Brown BlackBlue 0.75
- Page 115 and 116: 113Appendix CAppendix for Section 6
- Page 117 and 118: 115Appendix DPublicationsThe featur
- Page 119 and 120: 117Bibliography[AAR04] S. Agarwal,
- Page 121 and 122: 119[FCB08] L. Franssen, J. E. Coppe
- Page 123 and 124: 121[Ley96] M. Leyton. The architect
- Page 125 and 126: 123[RN11] D. Reid and M. Nixon. Usi
- Page 127 and 128: 125[ZG09] X. Zhang and Y. Gao. Face
- Page 129: 2Rapporteurs:Prof. Dr. Abdenour HAD
- Page 132 and 133: Biométrie faciale douce 2Les terme
- Page 134 and 135: Biométrie faciale douce 4une perso
- Page 136 and 137: Couleur depeauCouleur descheveuxCou
- Page 138 and 139: Biométrie faciale douce 8Nous nous
- Page 140 and 141: Biométrie faciale douce 103. Proba
- Page 142 and 143: Biométrie faciale douce 12l’entr
- Page 144 and 145: Biométrie faciale douce 14Figure 6
- Page 146 and 147: Biométrie faciale douce 16pages 77
- Page 148 and 149: Reviewers:Prof. Dr. Abdenour HADID,
- Page 150 and 151: 3hair, skin and clothes. The propos
- Page 152 and 153: person in the red shirt”. Further
102 8. USER ACCEPTANCE STUDY RELATING TO <strong>SOFT</strong> <strong>BIOMETRICS</strong>Applying s<strong>of</strong>t biometrics for person re-identification, with a focus on the frontal vs.side scenarioMotivated by realistic surveillance scenarios, the work addressed the problem <strong>of</strong> frontal-tosidefacial recognition, providing re–identification algorithms/classifiers that are specifically suitedfor this setting. Emphasis was placed on classifiers that belong in the class <strong>of</strong> s<strong>of</strong>t biometric traits,specifically color–, texture– and intensity– based traits taken from patches <strong>of</strong> hair, skin and clothes.Towards providing insight, the work presented different identification experiments that adhere tothe frontal–to–side setting, as well as presented a preliminary analytical study that seeks to impartintuition on the role <strong>of</strong> the above traits in improving algorithmic reliability. Our analysis describedthe overall error probability, both as a function <strong>of</strong> collisions and <strong>of</strong> erroneous categorizations forgiven sizes <strong>of</strong> authentication groups. In the presence <strong>of</strong> a moderate reliability <strong>of</strong> the patches-basedmethod, the analysis suggests promising applications <strong>of</strong> this method in settings such as pruning <strong>of</strong>searches.Applying s<strong>of</strong>t biometrics quantification and prediction <strong>of</strong> female facial aestheticsIn terms <strong>of</strong> female facial aesthetics, we presented a study on facial aesthetics in photographs,where we compared objective measures (namely photograph quality measures, facial beauty characteristicsand non permanent facial features), with human subjective perception. Our analysisrevealed a substantial correlation between different selected traits, and the corresponding MOSrelatedbeauty indices. Specifically we presented that non permanent features can influence highlythe MOS, and based on our analysis we conclude that facial aesthetics in images can indeedbe substantially modifiable. With other words parameters such as the presence <strong>of</strong> makeup andglasses, the image quality as well as different image post–processing methods can significantlyaffect the resulting MOS. Furthermore we constructed a linear MOS–based metric which wassuccessfully employed to quantify beauty-index variations due to aging and surgery. Our workapplies towards building a basis for designing new image-processing tools that further automateprediction <strong>of</strong> aesthetics in facial images. Towards this we provided a simulation <strong>of</strong> an automaticprediction tool based on state-<strong>of</strong>-art classification algorithms and the designed MOS–predictionmetric.The above approaches were accompanied by a more practically oriented part where we designedan automatic s<strong>of</strong>t biometrics classification tool. Specifically we focused on eye, skin andhair color, as well as on the presence <strong>of</strong> beard, moustache and glasses.In terms <strong>of</strong> usability analysis, we presented a user study investigating the preference <strong>of</strong> a set<strong>of</strong> test participants on access methods, namely s<strong>of</strong>t biometrics, face, PIN and fingerprint basedaccess methods. This preference was evaluated generally in terms <strong>of</strong> usability measures, suchas ease <strong>of</strong> use, intuitiveness and log-in-speed. Furthermore two scenarios were hereby assessed,specifically personal computer access and entrance <strong>of</strong> a security lab in a crowded environment.The surprising outcome is that although all users were strongly biased towards the PIN basedverification method, by daily use, the biometric based options were overall equally or even significantlybetter rated than the PIN based system. Users appreciated the comfort, easiness and speed<strong>of</strong> modern technology. Specifically they favored the s<strong>of</strong>t biometrics system, due to the providedprivacy preservation and ease <strong>of</strong> use.