Meta tags:

Headings (most frequently used words):

selected, welcome, news, personal, repositories, academic, activities, contact, huy, nguyen,

Text of the page (most frequently used words):
and (14), the (12), was (11), our (9), accepted (9), detection (8), published (8), arxiv (7), huy (6), for (6), ieee (6), work (6), deepfake (6), were (6), implementation (5), paper (5), adversarial (5), papers (5), 2023 (4), two (4), 2024 (4), japan (4), nguyen (3), apsipa (3), asc (3), session (3), privacy (3), learning (3), media (3), access (3), generation (3), from (3), attacks (3), ijcb (3), about (3), university (3), tokyo (3), please (2), contact (2), special (2), multimedia (2), security (2), deep (2), 2020 (2), acm (2), journal (2), wacv (2), academic (2), runtime (2), github (2), selected (2), 2019 (2), icassp (2), 2021 (2), dataset (2), fake (2), master (2), machine (2), previously (2), 2022 (2), vein (2), other (2), infrared (2), with (2), degree (2), national (2), institute (2), informatics (2), researcher (2), professor (2), nii (2), intuitions (2), welcome (2), 2026, powered, fork, minimal, mistakes, academicpages, jekyll, sitemap, age, generative, models, clones, its, chair, tip, tifs, caa, jas, tomm, elsevier, prletters, eaai, eurasip, jivp, ieice, neurips, iclr, icml, cvpr, eccv, icme, acl, conferences, reviewer, activities, governance, aware, agent, https, com, honghuy127, aegis, public, aegisruntime, personal, repositories, early, segmentation, btas, iccv, openforensics, book, chapters, frontiers, handbook, digital, face, manipulation, deepfakes, morphing, premier, faces, biom, applying, protection, another, same, topic, cvprw, wifs, one, finger, recognition, systems, theory, non, robust, features, transferability, examples, image, purification, open, signal, processing, source, code, will, soon, collaborative, physics, based, attack, near, human, detector, doi, oscars, lab, measuring, similarities, between, computer, vision, tasks, pioneer, restoration, namely, cyber, vaccine, nlp, multimodal, out, context, aina, detecting, text, watermarking, llms, comparative, study, fine, grained, counting, methods, ekyc, icip, defense, against, physical, domain, spurious, feature, three, news, more, details, have, look, last, update, 2025, here, received, honors, program, vietnam, 2013, alliance, graduate, advanced, studies, sokendai, vnuhcm, science, currently, softbank, group, company, visiting, associate, project, assistant, 特任助教, research, interests, include, safety, plus, echizen, laboratory, nguyễn, hồng, vietnamese, orcid, google, scholar, linkedin, twitter, researchgate, follow, talks, publications, grants, projects,

Text of the page (random words):
welcome huy huy projects grants publications talks huy h nguyen researcher at sb intuitions tokyo japan follow tokyo japan researchgate twitter linkedin github google scholar orcid welcome i am huy h nguyen nguyễn hồng huy in vietnamese i am currently a researcher at sb intuitions a softbank group company and a visiting associate professor at the national institute of informatics nii japan previously i was a project assistant professor 特任助教 at echizen laboratory nii my research interests include safety security and privacy for machine learning plus fake media generation and detection i received my bs degree honors program from vnuhcm university of science vietnam in 2013 and my ph d degree from the graduate university for advanced studies sokendai in alliance with the national institute of informatics japan in 2022 for more details please have a look at my cv here last update 2025 05 25 selected news three deepfake papers were accepted to ijcb 2024 two papers spurious feature generation defense against physical adversarial attacks on infrared domain were accepted to icip 2024 two papers ekyc deepfake dataset and a comparative study of fine grained counting methods were published in ieee access a paper about watermarking for llms was accepted to icassp 2024 two nlp papers detecting academic ai text and multimodal out of context detection were accepted to aina 2024 our pioneer work on deepfake restoration namely cyber vaccine was published in ieee access our work on measuring the similarities between other computer vision tasks to deepfake detection was accepted to ijcb 2023 arxiv our collaborative paper on physics based adversarial attack on near infrared human detector with oscars lab at the university of tokyo was accepted to acm multimedia 2023 doi implementation our paper on adversarial image purification was accepted to ieee open journal of signal processing paper source code will be published soon two papers were accepted to wacv 2023 one is about master vein attacks on finger vein recognition systems arxiv the other is about the theory of the non robust features on the transferability of adversarial examples arxiv our work on applying adversarial machine learning for privacy protection was accepted to wifs 2022 another work on the same topic was previously published in cvprw 2021 our premier work on master faces was published in ijcb 2020 arxiv and ieee t biom arxiv our book chapters on deepfake detection were published in handbook of digital face manipulation and detection from deepfakes to morphing attacks implementation and frontiers in fake media generation and detection our paper on the openforensics dataset was published in iccv 2021 implementation our early work on deepfake detection and segmentation was published in icassp 2019 arxiv implementation and btas 2019 arxiv implementation selected personal repositories aegisruntime governance aware agent runtime https github com honghuy127 aegis runtime public academic activities reviewer conferences neurips iclr icml cvpr eccv wacv icme acl rr apsipa asc journal ieee access tip tifs ieee caa jas acm tomm elsevier prletters eaai eurasip jivp ieice session chair apsipa asc 2020 special session deep generative models for media clones and its detection apsipa asc 2023 special session multimedia security and privacy in the age of deep learning contact please contact me at sitemap 2026 huy h nguyen powered by jekyll academicpages a fork of minimal mistakes