{"id":688,"date":"2016-03-15T09:54:00","date_gmt":"2016-03-15T08:54:00","guid":{"rendered":"https:\/\/samovar2022.int-evry.fr\/index.php\/2016\/03\/15\/oct-en-phase-pour-la-reconnaissance-biometrique-par-empreinte-digitale-et-sa-securisation\/"},"modified":"2020-09-04T18:46:32","modified_gmt":"2020-09-04T16:46:32","slug":"oct-en-phase-pour-la-reconnaissance-biometrique-par-empreinte-digitale-et-sa-securisation","status":"publish","type":"post","link":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/2016\/03\/15\/oct-en-phase-pour-la-reconnaissance-biometrique-par-empreinte-digitale-et-sa-securisation\/","title":{"rendered":"\u00ab\u00a0OCT en phase pour la reconnaissance biom\u00e9trique par empreinte digitale et sa s\u00e9curisation\u00a0\u00bb"},"content":{"rendered":"<p>Soutenance de th\u00e8se de <strong>Fran\u00e7ois LAMARE<\/strong><br \/>\n<strong>le 21 mars 2016 \u00e0 14h00 &#8211; Salle G10<\/strong> \u00e0 T\u00e9l\u00e9com SudParis, 9 rue Charles Fourier &#8211; 91011 Evry cedex.<\/p>\n<p>Cette th\u00e8se a \u00e9t\u00e9 r\u00e9alis\u00e9e sous la direction du Professeur <strong>Badr-Eddine BENKELFAT.<\/strong><\/p>\n<p>Le jury sera compos\u00e9 de :<\/p>\n<ul>\n<li>BEN AMOR Boulbaba Enseignant chercheur (HDR) Rapporteur<\/li>\n<li>TRIA Assia Enseignant chercheur (HDR) Rapporteur<\/li>\n<li>ZAQUINE Isabelle Professeur (HDR) Examinateur <\/li>\n<li>Liming CHEN Professeur Examinateur <\/li>\n<li>FOURRE Jo\u00ebl-Yann Directeur technique capteurs &#8211; Morpho Invit\u00e9<\/li>\n<li>GOTTESMAN Yaneck Maitre de conf\u00e9rences (HDR) Encadrant<\/li>\n<li>DORIZZI Bernadette Professeur (HDR) Directrice de th\u00e8se<\/li>\n<li>BENKELFAT Badr-Eddine Professeur (HDR) Directeur de th\u00e8se<\/li>\n<\/ul>\n<p><strong>R\u00e9sum\u00e9 :<\/strong><\/p>\n<p>Au sein de la soci\u00e9t\u00e9 actuelle, le besoin de conna\u00eetre l\u2019\u2019identit\u00e9 des personnes, que ce soit \u00e0 des fins d\u2019authentification ou de s\u00e9curisation, est devenu primordial. De par sa fiabilit\u00e9 et son aspect a priori tr\u00e8s s\u00e9curis\u00e9, la biom\u00e9trie s\u2019est impos\u00e9e au fil des ann\u00e9es comme une solution pertinente. La modalit\u00e9 biom\u00e9trique la plus largement r\u00e9pandue est l&#8217;empreinte digitale. Les capteurs standards permettent l&rsquo;acquisition d&rsquo;images d\u2019intensit\u00e9 en 2 dimensions de la surface du doigt, d\u2019o\u00f9 sont extraits des points caract\u00e9ristiques discriminants, les minuties. Les outils utilis\u00e9s pour l\u2019authentification biom\u00e9triques sont alors bas\u00e9s sur le r\u00e9sultat de la mise en correspondance (matching) des minuties de deux images, et cela suivant des propri\u00e9t\u00e9s d\u2019invariance par rotation, notamment. Ces outils peuvent \u00eatre consid\u00e9r\u00e9s \u00e0 l\u2019heure actuelle comme performants dans de bonnes conditions d\u2019acquisition.<\/p>\n<p>Le niveau de s\u00e9curit\u00e9 d\u2019un syst\u00e8me biom\u00e9trique d\u00e9pend entre autres des vuln\u00e9rabilit\u00e9s du capteur qui est utilis\u00e9 lors de l\u2019\u00e9tape d\u2019authentification. L\u2019identit\u00e9 d\u2019un individu peut en effet \u00eatre facilement usurp\u00e9e (spoofing) \u00e0 l\u2019aide de moyens tr\u00e8s simples, en faisant des attaques directes sur le capteur avec des faux doigts ou de fausses empreintes, confectionn\u00e9s \u00e0 partir des empreintes de cet individu. Pour certains syst\u00e8mes, ces failles de s\u00e9curit\u00e9 peuvent poser probl\u00e8me, comme par exemple dans le cas d\u2019une transaction bancaire sensible, ou bien d\u2019un acc\u00e8s \u00e0 une zone s\u00e9curis\u00e9e, ou encore d\u2019un contr\u00f4le du passage \u00e0 une fronti\u00e8re, etc&#8230; A travers ces exemples, nous comprenons tout l\u2019int\u00e9r\u00eat que repr\u00e9sentent ces failles de s\u00e9curit\u00e9 pour un fraudeur. Un autre probl\u00e8me important en biom\u00e9trie des empreintes digitales concerne des d\u00e9gradations possibles de la qualit\u00e9 des images obtenues. Ce ph\u00e9nom\u00e8ne peut \u00eatre li\u00e9 au syst\u00e8me d\u2019acquisition, qui peut par exemple \u00eatre sensible \u00e0 l\u2019\u00e9tat de surface de la peau (doigts secs ou mouill\u00e9s, par exemple), \u00e0 l\u2019environnement (notamment la temp\u00e9rature) ou voire tout simplement \u00e0 des d\u00e9gradations superficielles de l\u2019\u00e9piderme qui peuvent diminuer les performances de reconnaissance.<\/p>\n<p>D\u2019apr\u00e8s l\u2019analyse que nous faisons de cette situation, ces probl\u00e8mes sont dus essentiellement \u00e0 la faible quantit\u00e9 d&rsquo;information (information de surface uniquement), enregistr\u00e9e par les capteurs usuels durant la phase d\u2019authentification, et \u00e0 la nature physique de cette information (intensit\u00e9 seulement). Pour y rem\u00e9dier, nous proposons dans ce travail de th\u00e8se une nouvelle approche bas\u00e9e sur l\u2019OCT (\u00ab Optical Coherence Tomography \u00bb). L\u2019OCT est un capteur optique sans contact, bas\u00e9e sur une technique interf\u00e9rom\u00e9trique, qui permet d\u2019imager en 3 dimensions les \u00e9l\u00e9ments diffusants la lumi\u00e8re, et cela avec une haute r\u00e9solution. Il permet \u00e9galement d\u2019obtenir une information physique tr\u00e8s riche sur l\u2019\u00e9chantillon biom\u00e9trique analys\u00e9.<\/p>\n<p>L\u2019OCT est un capteur d\u00e9di\u00e9 initialement au monde m\u00e9dical (\u00e0 la base pour l\u2019ophtalmologie). La plupart des applications m\u00e9dicales reposent surtout sur des images d\u2019intensit\u00e9 en coupes transversales, pour le diagnostic m\u00e9dical par exemple. L\u2019originalit\u00e9 de notre travail consiste alors \u00e0 exploiter un tel instrument dans le contexte de la biom\u00e9trie, o\u00f9 l\u2019information pertinente est par essence tridimensionnelle (les empreintes digitales sont des surfaces 3D). L\u2019utilisation de l\u2019OCT dans un domaine diff\u00e9rent de son domaine d\u2019application originel pose alors un certain nombre de probl\u00e8mes quant \u00e0 la nature de l\u2019information exploit\u00e9e, mais aussi en termes de performances attendues et de dispositif instrumental.<\/p>\n<p>En compl\u00e9ment aux approches traditionnelles en OCT mais aussi en biom\u00e9trie, bas\u00e9es sur l\u2019information d\u2019intensit\u00e9, nous avons \u00e9tudi\u00e9 dans ce travail de th\u00e8se une autre modalit\u00e9 d\u2019imagerie OCT, bas\u00e9e sur l\u2019exploitation de la phase. Cette phase, directement li\u00e9e \u00e0 des mesures de temps de vol de la lumi\u00e8re, permet alors de r\u00e9pondre au probl\u00e8me de la segmentation des empreintes digitales 3D, telles qu\u2019obtenues dans un mode d\u2019acquisition sans contact (absence d\u2019aplatissement des doigts). Nous avons d\u2019ailleurs montr\u00e9 gr\u00e2ce \u00e0 des simulations statistiques, mod\u00e9lisant l\u2019impact du bruit du capteur sur les mesures de temps de vol, que ces surfaces 3D peuvent \u00eatre extraites du volume tomographique avec une grande pr\u00e9cision (). Nous avons \u00e9galement propos\u00e9 dans ce travail de th\u00e8se une repr\u00e9sentation originale des empreintes, dite en contraste de phase, qui permet d\u2019obtenir des images de tr\u00e8s bonne qualit\u00e9, m\u00eame en conditions d\u2019acquisitions d\u00e9grad\u00e9es (doigts humides), contrairement \u00e0 une repr\u00e9sentation plus conventionnelle en intensit\u00e9.<\/p>\n<p>L\u2019information 3D obtenue gr\u00e2ce \u00e0 l\u2019OCT ne se limite pas seulement \u00e0 l\u2019empreinte \u00e0 la surface du doigt. Il existe notamment une deuxi\u00e8me empreinte, dite empreinte interne, situ\u00e9e entre le derme et l\u2019\u00e9piderme, dont la structure est tr\u00e8s similaire \u00e0 celle de l\u2019empreinte surfacique traditionnelle (dite empreinte externe). Nous nous sommes particuli\u00e8rement int\u00e9ress\u00e9s \u00e0 cette empreinte interne dans notre travail de th\u00e8se. En effet, cette derni\u00e8re est plus difficile d&rsquo;acc\u00e8s que l\u2019empreinte externe, ce qui est d&rsquo;un grand int\u00e9r\u00eat pour la s\u00e9curit\u00e9 (anti-spoofing), mais est \u00e9galement mieux pr\u00e9serv\u00e9e des conditions ext\u00e9rieures (humidit\u00e9, salet\u00e9s, blessures, \u2026). La contrepartie de cet avantage est qu\u2019elle est alors beaucoup plus difficile \u00e0 extraire (\u00e0 cause de sa profondeur, mais aussi des processus de diffusion de la lumi\u00e8re \u00e0 l\u2019int\u00e9rieur des tissus biologiques). En particulier, il s\u2019est av\u00e9r\u00e9 que l\u2019utilisation de notre repr\u00e9sentation en contraste de phase \u00e9tait insatisfaisante pour l\u2019obtention d\u2019images d\u2019empreintes internes de bonne qualit\u00e9. Une autre contribution importante de ce travail de th\u00e8se repose alors sur un nouveau proc\u00e9d\u00e9 d&rsquo;imagerie, bas\u00e9 sur la fusion de l\u2019information en phase et en intensit\u00e9, qui sont compl\u00e9mentaires dans le cas des empreintes internes. Notre m\u00e9thode de fusion est bas\u00e9e sur des mesures de la qualit\u00e9 locale des images d\u2019empreintes dans les deux repr\u00e9sentations. Elle permet de ne conserver que les zones de meilleure qualit\u00e9 issues de l&rsquo;une ou l&rsquo;autre de ces repr\u00e9sentations. Notre m\u00e9thode d\u2019imagerie par fusion permet donc d\u2019obtenir des images de bien meilleure qualit\u00e9 que celle des images en intensit\u00e9 et en contraste de phase prises isolement.<\/p>\n<p>Nos approches en phase et en fusion ont pu \u00eatre \u00e9valu\u00e9es \u00e0 partir d\u2019exp\u00e9riences en v\u00e9rification d\u2019empreintes (bas\u00e9es sur des matching de minuties), effectu\u00e9es sur une petite base de donn\u00e9es d\u2019empreintes digitales 3D, acquise au sein de notre laboratoire (une centaine de doigts diff\u00e9rents). R\u00e9aliser le matching de surfaces 3D pour l\u2019authentification est de nos jours une t\u00e2che encore difficile \u00e0 mettre en \u0153uvre. Nous nous sommes alors ramen\u00e9es au cas d\u2019images 2D, pour lequel il existe des outils de matching d\u00e9j\u00e0 performants. Nous avons donc \u00e0 cette fin utilis\u00e9e une m\u00e9thode de projection issue de l\u2019\u00e9tat de l\u2019art, que nous avons adapt\u00e9 au cas des empreintes digitales. La m\u00e9thode repose sur le paradigme de la minimisation des distorsions des distances g\u00e9od\u00e9siques lors de l\u2019aplatissement, particuli\u00e8rement judicieux dans un contexte biom\u00e9trique o\u00f9 la conservation des distances est capitale. Par ailleurs, nous avons montr\u00e9 que la m\u00e9thode employ\u00e9e respectait la propri\u00e9t\u00e9 importante d\u2019invariance \u00e0 la rotation des doigts. Les diff\u00e9rents r\u00e9sultats obtenus \u00e0 partir des images d\u2019empreintes aplaties ont alors permis de valider l\u2019int\u00e9r\u00eat de notre repr\u00e9sentation en contraste de phase par rapport \u00e0 celle en intensit\u00e9, dans le cas de l\u2019empreinte externe, mais \u00e9galement de montrer la pertinence de notre m\u00e9thode de fusion, dans le cas de l\u2019empreinte interne. Les performances en authentification obtenues avec notre approche globale OCT ont \u00e9galement \u00e9t\u00e9 compar\u00e9es \u00e0 celles obtenues avec des capteurs 2D usuels, d\u00e9j\u00e0 largement optimis\u00e9s en termes de taux d\u2019erreurs. Les r\u00e9sultats obtenus montrent que l\u2019OCT est une approche tout \u00e0 fait prometteuse pour l\u2019authentification d\u2019individus.<\/p>\n<p>Enfin, nous avons adress\u00e9 dans ce travail de th\u00e8se le probl\u00e8me du spoofing, particuli\u00e8rement critique dans le cas des empreintes digitales. Nous avons montr\u00e9 que l\u2019utilisation d&rsquo;un capteur OCT \u00e9tait tr\u00e8s int\u00e9ressante dans la mesure o\u00f9 l\u2019information obtenue, de profondeur notamment, pouvait \u00eatre exploit\u00e9e avantageusement pour d\u00e9tecter la pr\u00e9sence d\u2019un faux doigt ou d\u2019une fausse empreinte. Ainsi, la fabrication d\u2019un leurre parfaitement ind\u00e9tectable par l\u2019OCT est rendu particuli\u00e8rement difficile, compliquant grandement le travail d\u2019un faussaire. Nous avons \u00e9galement propos\u00e9 une nouvelle approche tr\u00e8s prometteuse pour l\u2019anti-spoofing, qui consiste \u00e0 comparer les diff\u00e9rentes images d&#8217;empreintes d\u2019un m\u00eame doigt. Dans le cas d\u2019une usurpation d\u2019identit\u00e9, l\u2019empreinte \u00e0 la surface (correspondant donc au leurre) serait diff\u00e9rente des empreintes externe (empreinte sur le doigt) et interne (empreinte dans le doigt). Nous serions alors en mesure de d\u00e9tecter la tentative de fraude, et aussi utiliser l\u2019empreinte de surface pour de d\u00e9terminer quelle identit\u00e9 a \u00e9t\u00e9 usurp\u00e9e (\u00e0 condition que cette empreinte soit connue bien entendu). Nous voyons ici encore l\u2019int\u00e9r\u00eat de nos m\u00e9thodes en phase et en fusion, et aussi de notre m\u00e9thode d\u2019aplatissement, pour obtenir ces diff\u00e9rentes images d\u2019empreintes mais aussi pour les comparer.<\/p>\n<p><strong>Abstract :<\/strong><\/p>\n<p>In today&rsquo;s society, the need to know the identity of persons, whether for authentication or security purposes, has become primordial. Because of its reliability and its appearance a priori very secure, biometrics has emerged over the years as a relevant solution.<\/p>\n<p>Fingerprint is currently the most widespread biometric modality. The standard sensors allow the acquisition of intensity images in 2 dimensions of the finger surface, where are extracted discriminating points called minutiae. The tools used for biometric authentication are then based on the result of matching of the minutiae of two images, with a rotation invariance property, in particular. These tools can be considered at present as efficient in good conditions of acquisition.<\/p>\n<p>The security level of a biometric system inter alia depends on the sensor vulnerabilities that is used during the authentication step. The identity of an individual can indeed be easily spoofed using very simple means, making direct attacks on the sensor with fake fingers or false prints, made from impressions of this individual. For some systems, these security lacks can cause problems, such as in the case of sensitive banking transaction, or access to a secure area, or of control of crossing a border, etc. Through these examples, we understand the interest represented by these security problems to a fraudster. Another important problem in fingerprint biometrics relates to possible degradation of the quality of the obtained images. This can be linked to the acquisition system, which may for instance be sensitive to the skin surface condition (dry or wet fingers, for example), the environment (such as temperature) or even simply superficial damages to the skin that can decrease recognition performance.<\/p>\n<p>The analysis we make of this situation is that these problems are mainly due to the small amount of information (surface information only) registered by the standards sensors during the authentication phase, and the physical nature of this information (intensity only). To solve this, we propose in this thesis a new approach based on the OCT (\u00ab\u00a0Optical Coherence Tomography\u00a0\u00bb). OCT is a contactless optical sensor, based on an interferometric technique, which allows imaging in 3-dimensions the light scattering elements, and this with an high resolution. It also provides a wealth of information on physical biometric sample analyzed.<\/p>\n<p>OCT is a sensor initially dedicated to the medical world (ophthalmology at beginning). Most medical applications are primarily based on intensity cross-sectional images, for example for medical diagnosis. The originality of our work is then to use such an instrument in the context of biometrics, where the relevant information is three-dimensional (fingerprints is a 3D surface). The use of OCT in a different area of its original area of application raises a number of issues about the nature of the information exploited, but also in terms of expected performance and instrumental device.<\/p>\n<p>In addition to traditional approaches in OCT but also in biometrics, based on intensity information, we have studied in this thesis another imaging modality, based on the phase. This phase, directly related to light time of flight measurements, then allows responding to the problem of segmenting fingerprints in 3D, as obtained in an contactless acquisition (no flattening of the fingers ). We showed through statistical simulations, modeling the impact of sensor noise on time of flight measurements, than these 3D surfaces can be extracted from the tomographic volume with high precision (~ \u03bb \/ 10). We also proposed in this thesis an original representation of fingerprints, called phase contrast, which allows obtaining images of very good quality, even in degraded conditions acquisitions (wet fingers), unlike a conventional intensity representation.<\/p>\n<p>The 3D information obtained through OCT is not only limited to fingerprint at finger surface. In particular, there is a second fingerprint, called internal fingerprint, between the dermis and the epidermis, whose structure is very similar to the traditional fingerprint at the finger surface (called external fingerprint). We are particularly interested in this internal fingerprint in the thesis. Indeed, the latter is more difficult to access than the external fingerprint, which is of great interest for security (anti-spoofing), but is also better preserved from external conditions (humidity, dirt, damage, &#8230; ). The counterpart of this advantage is that it is much more difficult to extract (because of its depth, but also the process of diffusion of light inside biological tissue). In particular, it has been found that the use of our representation phase contrast was unsatisfactory for obtaining internal fingerprint images of good quality. Another important contribution of this thesis is based on a new imaging method based on of phase and intensity fusion information, which are complementary in the case of internal fingerprints. Our fusion method is based on measurements of the local quality of fingerprint images in the two representations. It allows keeping only the best quality areas from one or other of these representations. Our fusion imaging method allows obtaining images of higher quality than images in intensity and phase contrast, taken isolation. It has been the subject of a patent pending.<\/p>\n<p>Our phase and fusion approaches have been be evaluated with fingerprint verification experiments (based on matching of minutiae), realized on a small base of 3D fingerprint data acquired in our laboratory (a hundred different fingers). Nowadays, achieve 3D surfaces matching for authentication is a task still difficult to implement. We then reduced to the case of 2D images for which there are already performing matching tools. We therefore proposed to use at this end a projection method of the stat of the art that we have adapted to the case of fingerprints. The method is based on the paradigm of minimizing distortions of geodesic distances when flattening, particularly appropriate in a biometrical context where the distances conservation is mandatory. Furthermore, we have shown that the method respects the important property of invariance to rotation of the fingers. The different results obtained from the flattened fingerprint images were then used to validate the benefit of our phase contrast representation relative to the intensity one in the case of the external fingerprint, but also to show the relevance of our fusion imaging method, in the case of the internal fingerprint. The authentication performances obtained with our OCT approach were also compared to those obtained with conventional 2D sensors, already widely optimized in terms of error rates. The results obtained show that OCT is a promising approach for authentication of individuals.<\/p>\n<p>Finally, we have addressed in this thesis the problem of spoofing, particularly critical in the case of fingerprints. We have shown that the use of an OCT sensor was very interesting insofar 3D information and physical obtained can be advantageously exploited to detect the presence of a false finger or a false fingerprint. Thus, producing a perfectly undetectable fake is made particularly difficult, greatly complicating the work of a fraudster. We also proposed a promising new approach for anti-spoofing, which involves comparing the various images of fingerprints of the same finger. In the case of identity theft, the fingerprint on the surface (thus corresponding to the fake fingerprint) is different from the external (on the finger) and internal (inside the finger) fingerprints. We would then be able to detect attempted fraud, but also use the surface fingerprint to determine which identity is misused (provided that this fingerprint is known of course). We see here again the interest of our phase and fusion imaging methods, and also our flattening method to obtain these different fingerprint images, and also to compare them.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Soutenance de th\u00e8se de Fran\u00e7ois LAMARE le 21 mars 2016 \u00e0 14h00 &#8211; Salle G10 \u00e0 T\u00e9l\u00e9com SudParis, 9 rue Charles Fourier &#8211; 91011 Evry cedex. Cette th\u00e8se a \u00e9t\u00e9 r\u00e9alis\u00e9e sous la direction du Professeur Badr-Eddine BENKELFAT. Le jury sera compos\u00e9 de : BEN AMOR Boulbaba Enseignant chercheur (HDR) Rapporteur TRIA Assia Enseignant chercheur [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ocean_post_layout":"","ocean_both_sidebars_style":"","ocean_both_sidebars_content_width":0,"ocean_both_sidebars_sidebars_width":0,"ocean_sidebar":"","ocean_second_sidebar":"","ocean_disable_margins":"enable","ocean_add_body_class":"","ocean_shortcode_before_top_bar":"","ocean_shortcode_after_top_bar":"","ocean_shortcode_before_header":"","ocean_shortcode_after_header":"","ocean_has_shortcode":"","ocean_shortcode_after_title":"","ocean_shortcode_before_footer_widgets":"","ocean_shortcode_after_footer_widgets":"","ocean_shortcode_before_footer_bottom":"","ocean_shortcode_after_footer_bottom":"","ocean_display_top_bar":"default","ocean_display_header":"default","ocean_header_style":"","ocean_center_header_left_menu":"","ocean_custom_header_template":"","ocean_custom_logo":0,"ocean_custom_retina_logo":0,"ocean_custom_logo_max_width":0,"ocean_custom_logo_tablet_max_width":0,"ocean_custom_logo_mobile_max_width":0,"ocean_custom_logo_max_height":0,"ocean_custom_logo_tablet_max_height":0,"ocean_custom_logo_mobile_max_height":0,"ocean_header_custom_menu":"","ocean_menu_typo_font_family":"","ocean_menu_typo_font_subset":"","ocean_menu_typo_font_size":0,"ocean_menu_typo_font_size_tablet":0,"ocean_menu_typo_font_size_mobile":0,"ocean_menu_typo_font_size_unit":"px","ocean_menu_typo_font_weight":"","ocean_menu_typo_font_weight_tablet":"","ocean_menu_typo_font_weight_mobile":"","ocean_menu_typo_transform":"","ocean_menu_typo_transform_tablet":"","ocean_menu_typo_transform_mobile":"","ocean_menu_typo_line_height":0,"ocean_menu_typo_line_height_tablet":0,"ocean_menu_typo_line_height_mobile":0,"ocean_menu_typo_line_height_unit":"","ocean_menu_typo_spacing":0,"ocean_menu_typo_spacing_tablet":0,"ocean_menu_typo_spacing_mobile":0,"ocean_menu_typo_spacing_unit":"","ocean_menu_link_color":"","ocean_menu_link_color_hover":"","ocean_menu_link_color_active":"","ocean_menu_link_background":"","ocean_menu_link_hover_background":"","ocean_menu_link_active_background":"","ocean_menu_social_links_bg":"","ocean_menu_social_hover_links_bg":"","ocean_menu_social_links_color":"","ocean_menu_social_hover_links_color":"","ocean_disable_title":"default","ocean_disable_heading":"default","ocean_post_title":"","ocean_post_subheading":"","ocean_post_title_style":"","ocean_post_title_background_color":"","ocean_post_title_background":0,"ocean_post_title_bg_image_position":"","ocean_post_title_bg_image_attachment":"","ocean_post_title_bg_image_repeat":"","ocean_post_title_bg_image_size":"","ocean_post_title_height":0,"ocean_post_title_bg_overlay":0.5,"ocean_post_title_bg_overlay_color":"","ocean_disable_breadcrumbs":"default","ocean_breadcrumbs_color":"","ocean_breadcrumbs_separator_color":"","ocean_breadcrumbs_links_color":"","ocean_breadcrumbs_links_hover_color":"","ocean_display_footer_widgets":"default","ocean_display_footer_bottom":"default","ocean_custom_footer_template":"","ocean_post_oembed":"","ocean_post_self_hosted_media":"","ocean_post_video_embed":"","ocean_link_format":"","ocean_link_format_target":"self","ocean_quote_format":"","ocean_quote_format_link":"post","ocean_gallery_link_images":"on","ocean_gallery_id":[],"footnotes":""},"categories":[350],"tags":[],"class_list":["post-688","post","type-post","status-publish","format-standard","hentry","category-theses-2016-fr","entry"],"_links":{"self":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/688","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/comments?post=688"}],"version-history":[{"count":1,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/688\/revisions"}],"predecessor-version":[{"id":1700,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/688\/revisions\/1700"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=688"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=688"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=688"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}