https:\/\/webconf.imt.fr\/frontend\/rooms\/kea-bjw-vkl-gru\/join<\/a><\/p>\n\n\n\nMembres du jury :<\/strong><\/p>\n\n\n\nM. Catalin FETITA<\/strong>, Professeur, T\u00e9l\u00e9com SudParis, FRANCE – Directeur de these
Mme Elisabeth BRUNET<\/strong>, Ma\u00eetresse de conf\u00e9rences, T\u00e9l\u00e9com SudParis, FRANCE – Co-encadrante de these
M. Pierre FORTIN<\/strong>, Professeur des universit\u00e9s, Universit\u00e9 de Lille , FRANCE – Rapporteur
M. Pascal DESBARATS<\/strong>, Professeur, Universit\u00e9 de Bordeaux, FRANCE – Rapporteur
M. Raymond NAMYST<\/strong>, Professeur, Universit\u00e9 de Bordeaux, FRANCE – Examinateur
Mme Maria VAKALOPOULOU<\/strong>, Ma\u00eetresse de conf\u00e9rences, Centrale Sup\u00e9l\u00e9c, FRANCE – Examinateur<\/p>\n\n\n\nInvit\u00e9s :<\/strong><\/p>\n\n\n\nMme Rosalie MARTIN<\/strong>, Ing\u00e9nieure de recherche, Adobe<\/p>\n\n\n\nM. Rafael PIRES<\/strong>, Enseignant-Chercheur, Ecole Polytechnique F\u00e9d\u00e9rale de Lausanne<\/p>\n\n\n\nM. Daniel Stockholm<\/strong>, Ma\u00eetre de Conf\u00e9rence HDR, EPHE-PSL (CRSA) <\/p>\n\n\n\n\u00ab Analyse multi-r\u00e9solution frugale pour les images en gigapixels: application aux donn\u00e9es biom\u00e9dicales et au-del\u00e0 \u00bb<\/h2>\n\n\n\npr\u00e9sent\u00e9 par Madame Marie REINBIGLER<\/h2>\n\n\n\n
R\u00e9sum\u00e9 :<\/strong><\/p>\n\n\n\nLes avanc\u00e9es r\u00e9alis\u00e9es dans le domaine des syst\u00e8mes d’acquisition d’images ont permis de capturer des images atteignant l’\u00e9chelle du milliard de pixels et comprenant un niveau de d\u00e9tail sans pr\u00e9c\u00e9dent, ce qui a consid\u00e9rablement accru les possibilit\u00e9s d’analyse dans une grande diversit\u00e9 de domaines. En particulier, le domaine biom\u00e9dical s’appuie sur des images en gigapixels de coupes de tissus, appel\u00e9es images histologiques, pour r\u00e9aliser des diagnostics qui constituent aujourd’hui la r\u00e9f\u00e9rence pour la d\u00e9tection du cancer. Cependant, l’analyse de ces images reste une t\u00e2che complexe, chronophage et sujette aux erreurs pour les praticiens. L’automatisation de ces t\u00e2ches est n\u00e9cessaire mais difficile, car la taille des images d\u00e9passe souvent les ressources disponibles. Ce probl\u00e8me est renforc\u00e9 par les besoins en ressources informatiques croissantes pour l’analyse d’images reposant sur des mod\u00e8les d’Intelligence Artificielle progressivement adopt\u00e9s. Les ressources informatiques requises ne sont pas toujours accessibles pour les praticiens et ont un impact socio-environnemental non n\u00e9gligeable, en particulier lorsqu’elles impliquent la fabrication de nouveaux composants. Il est donc primordial d’optimiser l’utilisation des \u00e9quipements informatiques existants. Le domaine de l’informatique frugale constitue une option permettant d’utiliser des ordinateurs plus anciens ou plus modestes pour r\u00e9aliser de telles analyses. Pour relever ces d\u00e9fis, nous proposons une approche \u00e9conome en calcul pour l’analyse d’images en gigapixels en exploitant leur structure multi-r\u00e9solution. Nous \u00e9valuons la m\u00e9thode dans deux sc\u00e9narios : l’un avec une distribution h\u00e9t\u00e9rog\u00e8ne des donn\u00e9es pertinentes, illustrant l’efficacit\u00e9 de l’approche, et l’autre avec une distribution spatiale homog\u00e8ne des zones d’int\u00e9r\u00eat, o\u00f9 la m\u00e9thode introduit un surco\u00fbt. Pour poursuivre notre qu\u00eate d’une ex\u00e9cution plus frugale, nous d\u00e9ployons notre approche sur une grappe d’ordinateurs standards et modestes, en s\u00e9lectionnant des strat\u00e9gies d’\u00e9quilibrage de charge adapt\u00e9es \u00e0 la densit\u00e9 d’informations de chaque cas d’utilisation. Enfin, nous d\u00e9montrons la g\u00e9n\u00e9ralisabilit\u00e9 de notre approche dans d’autres domaines, telles la d\u00e9tection de feu de for\u00eat \u00e0 l’aide d’images satellites et la simulation multi-\u00e9chelle des \u00eelots de chaleur urbains.<\/p>\n\n\n\n
Abstract :<\/strong><\/p>\n\n\n\nAdvancements in image acquisition systems have made it possible to capture gigapixel-scale images with an unprecedented level of detail, significantly reshaping analysis possibilities across a wide range of domains. In particular, the biomedical domain relies on gigapixel images of tissue slices, called histological images, to perform diagnosis and constitute the gold standard for cancer detection. However, analyzing these images remains a demanding, time-consuming, and error-prone task for pathologists. Automating these tasks is necessary but challenging, as the large image size often exceeds available resources. This issue is reinforced by the increasing computational demands of Artificial Intelligence-based image analysis algorithms that are progressively adopted. The required computation-heavy resources are not always affordable for pathologists and have a non-negligible socio-environmental impact, especially when it implies new device manufacturing. It is thus paramount to optimize the utilization of existing computing hardware. The frugal computing domain constitutes a direction to allow the use of older or more modest computers to achieve such analyses. To address these challenges, we propose a computation-saving approach for analyzing gigapixel images by exploiting their multi-resolution structure. We evaluate the method in two scenarios: one with a heterogeneous distribution of relevant data, demonstrating the approach efficiency, and another with a spatially homogeneous distribution, where the method introduces an overhead. To pursue our quest for more frugal computing, we implement our approach on a cluster of standard and modest computers, selecting load-balancing strategies tailored to the information density of each use case. Finally, we highlight the broader applicability of our method in other domains, like fire detection using satellite imagery and Urban Heat Island multi-scale simulation.<\/p>\n","protected":false},"excerpt":{"rendered":"
L’Ecole doctorale : Ecole Doctorale de l’Institut Polytechnique de Paris et le Laboratoire de recherche SAMOVAR – Services r\u00e9partis, Architectures, Mod\u00e9lisation, Validation, Administration des R\u00e9seaux pr\u00e9sentent l\u2019AVIS DE SOUTENANCE de Madame Marie REINBIGLER Autoris\u00e9e \u00e0 pr\u00e9senter ses travaux en vue de l\u2019obtention du Doctorat de l’Institut Polytechnique de Paris, pr\u00e9par\u00e9 \u00e0 T\u00e9l\u00e9com SudParis en : […]<\/p>\n","protected":false},"author":4,"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":[286,169],"tags":[],"class_list":["post-6889","post","type-post","status-publish","format-standard","hentry","category-fractualites-ennews-fr","category-seminaires-armedia","entry"],"_links":{"self":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6889","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\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/comments?post=6889"}],"version-history":[{"count":1,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6889\/revisions"}],"predecessor-version":[{"id":6890,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6889\/revisions\/6890"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=6889"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=6889"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=6889"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}