pr\u00e9sentent<\/p>\n\n\n\n
Autoris\u00e9 \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\n\n\n
le MARDI 30 SEPTEMBRE 2025 \u00e0 10h00<\/p>\n\n\n\n
\u00e0<\/p>\n\n\n\n
Amphith\u00e9\u00e2tre 11
T\u00e9l\u00e9com SudParis 9 rue Charles Fourier 91011 Evry Cedex<\/p>\n\n\n\n
Membres du jury :<\/strong><\/p>\n\n\n\n Mme Sonia GARCIA-SALICETTI<\/strong>, Professeure, T\u00e9l\u00e9com SudParis, FRANCE – Directeur de these R\u00e9sum\u00e9 :<\/strong><\/p>\n\n\n\n La marche humaine est une activit\u00e9 motrice complexe, qui se d\u00e9veloppe progressivement et devient de plus en plus vuln\u00e9rable au vieillissement et aux pathologies neurologiques telles que l\u2019AVC ou la maladie de Parkinson (MP). Ces troubles entra\u00eenent souvent une faiblesse musculaire ou une paralysie, et leur pr\u00e9valence croissante accentue le besoin d\u2019outils pr\u00e9cis d\u2019\u00e9valuation de la marche. Bien que l\u2019Analyse Quantifi\u00e9e de la Marche fournisse des mesures objectives via la capture de mouvement et la cin\u00e9matique articulaire, la pratique clinique repose encore largement sur des scores observationnels ou des indices quantitatifs simplifi\u00e9s, souvent insuffisants pour refl\u00e9ter la variabilit\u00e9 et l\u2019asym\u00e9trie dynamique de la marche. Cette th\u00e8se vise \u00e0 d\u00e9velopper des m\u00e9thodes quantitatives fines pour \u00e9valuer les d\u00e9viations et les asym\u00e9tries de la marche en exploitant la structure temporelle compl\u00e8te des s\u00e9quences cin\u00e9matiques angulaires. \u00c0 partir de bases de donn\u00e9es cliniques comprenant des sujets sains et des patients atteints de troubles moteurs (post-AVC et MP), nous proposons d\u2019abord une approche bas\u00e9e sur le signal, utilisant l\u2019algorithme K-Medoids et la distance Dynamic Time Warping (DTW), afin de capturer la variabilit\u00e9 du sch\u00e9ma de marche normal et de quantifier les \u00e9carts individuels. Pour d\u00e9passer les limites des comparaisons scalaires, nous introduisons un cadre d\u2019analyse bas\u00e9 sur l\u2019image via les Dissimilarity Maps (DM). Cela inclut les Deviation Dissimilarity Maps (DDM), pour une analyse riche spatio-temporelle des d\u00e9viations, ainsi que les Bilateral Dissimilarity Maps (BiDM), pour l\u2019\u00e9valuation d\u00e9taill\u00e9e de l\u2019asym\u00e9trie. Deux nouveaux indices cliniques (EGAI et GAD) sont d\u00e9riv\u00e9s de ces cartes \u00e0 l\u2019aide d\u2019une d\u00e9composition en valeurs singuli\u00e8res (SVD). Nous introduisons \u00e9galement les Self-Dissimilarity Maps (SDM), qui encodent la structure interne de chaque cycle de marche. Combin\u00e9es \u00e0 des autoencodeurs convolutionnels, elles permettent une \u00e9valuation globale robuste des d\u00e9viations dans un espace d\u2019embedding \u00e0 haute dimension. Enfin, nous d\u00e9veloppons un Visual Transformer Autoencoder (ViTAE) qui traite les SDMs sous forme de patchs. Cette m\u00e9thode fournit une analyse localis\u00e9e et cliniquement pertinente des d\u00e9viations, mettant en \u00e9vidence les instants pr\u00e9cis du cycle de marche associ\u00e9s \u00e0 la pathologie ou \u00e0 l\u2019effet th\u00e9rapeutique. Abstract :<\/strong><\/p>\n\n\n\n Human gait is a complex motor activity, progressively developed and increasingly affected by aging and neurological disorders such as stroke or Parkinson\u2019s disease (PD). These conditions often result in muscle weakness or paralysis, and with their rising prevalence, the need for precise gait assessment tools becomes critical. While Quantified Gait Analysis (QGA) offers objective insights via motion capture and joint kinematics, clinical practice still relies on observational or simplified quantitative scores that often overlook gait variability and dynamic asymmetries. This thesis aims to develop refined, quantitative methods for evaluating gait deviations and asymmetries by leveraging the full temporal structure of angular kinematic sequences. Using clinical datasets of healthy subjects and patients with motor impairments (post-stroke and PD), we first introduce a signal-based approach using K-Medoids and Dynamic Time Warping (DTW) to capture healthy gait variability and compute deviation scores per patient. To overcome the limitations of scalar comparisons, we propose a novel image-based framework using Dissimilarity Maps (DMs). This includes Deviation Dissimilarity Maps (DDMs) for rich spatiotemporal deviation analysis and Bilateral Dissimilarity Maps (BiDMs) for detailed asymmetry assessment, from which two new clinical indices (EGAI and GAD) are derived using Singular Value Decomposition. Additionally, we propose Self-Dissimilarity Maps (SDMs) that encode each gait cycle’s internal structure. Using convolutional autoencoders, these SDMs offer a robust global deviation score in a high-dimensional embedding space. Finally, to enable localized analysis, we develop a vision transformer autoencoder (ViTAE) that processes SDMs in patches. This method delivers localized and clinically meaningful information about gait deviations, highlighting precise moments within the gait cycle that relate to pathology or therapeutic effect.<\/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 Monsieur Lorenzo HERMEZ Autoris\u00e9 \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-6896","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\/6896","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=6896"}],"version-history":[{"count":1,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6896\/revisions"}],"predecessor-version":[{"id":6897,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6896\/revisions\/6897"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=6896"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=6896"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=6896"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
M. G\u00e9rard DRAY<\/strong>, Professeur, IMT Mines Al\u00e8s, FRANCE – Rapporteur
Mme Nesma HOUMANI<\/strong>, Ma\u00eetresse de conf\u00e9rences, T\u00e9l\u00e9com SudParis, FRANCE – Co-encadrant de these
M. Fran\u00e7ois ROUSSEAU<\/strong>, Professeur, IMT Atlantique, FRANCE – Rapporteur
M. J\u00e9rome BOUDY<\/strong>, Professeur, T\u00e9l\u00e9com SudParis, FRANCE – Examinateur
M. Anthony FLEURY<\/strong>, Professeur, IMT Nord Europe, FRANCE – Examinateur
M. Mathieu LEMPEREUR<\/strong>, Ing\u00e9nieur de recherche, CHRU de Brest, FRANCE – Examinateur<\/p>\n\n\n\n\u00ab Mod\u00e9lisation spatiotemporelle de la marche pour caract\u00e9riser les troubles moteurs par apprentissage automatique: du signal vers l\u2019analyse d\u2019image \u00bb<\/h2>\n\n\n\n
pr\u00e9sent\u00e9 par Monsieur Lorenzo HERMEZ<\/h2>\n\n\n\n
<\/p>\n\n\n\n