4,838 research outputs found
Angle- and spin-resolved photoelectron spectroscopy of the 4f subshell in atomic ytterbium
No full textSvensson A, Müller M, Böwering N, Heinzmann U, Radojevic V, Wijesundera W. Angle- and spin-resolved photoelectron spectroscopy of the 4f subshell in atomic ytterbium. Journal of Physics B: Atomic, Molecular and Optical Physics. 1988;21(8):L179-L185.The spin polarisation components of photoelectrons from atomic ytterbium have been measured over the photon energy range from 15.5 to 22.5 eV. The measurements were performed by making use of circularly polarised synchrotron radiation from the storage ring BESSY in conjunction with an angle-resolved electron spectrometer. Theoretical predictions based on the relativistic random-phase approximation (RRPA) are in fair agreement with the experimental data even though a certain offset between theory and experiment exists
Remote photoplethysmography : advancing robustness, privacy and security
No full textAbstract
Remote photoplethysmography (rPPG) offers a non-contact method for extracting physiological signals from facial videos, presenting a promising alternative to traditional contact-based techniques. This thesis addresses critical challenges encountered in rPPG, focusing on improving signal extraction robustness, ensuring privacy, and enhancing security against potential attacks.
For robust rPPG measurement, this thesis introduces novel self-supervised and transformer-based deep learning methods. Specifically, a long temporal context transformer network is developed, enhanced by a multi-task pre-training strategy, demonstrating state-of-the-art performance. Furthermore, fully self-supervised contrastive learning frameworks are proposed, incorporating comprehensive physiological priors and intra–inter data relationships, achieving performance that is competitive with supervised methods across multiple datasets.
To address privacy and security challenges, this thesis investigates both de-identification and attack vulnerability in rPPG systems. A learning-based facial video de-identification method is introduced, preserving rPPG signal integrity while significantly impairing biometric recognition. Additionally, two novel datasets are created to comprehensively study physical domain attacks and rPPG-based presentation attacks, revealing critical vulnerabilities in existing rPPG methods and highlighting the need for enhanced security measures.
In summary, this thesis contributes significantly to the field of rPPG by developing robust signal extraction algorithms, addressing privacy concerns through de-identification, and identifying and studying security vulnerabilities. These advancements pave the way for more reliable and secure rPPG applications in healthcare, security, and beyond. Original papers Savic, M., & Zhao, G. (2025). PhySU-Net: Long temporal context transformer for rPPG with self-supervised pre-training. Lecture Notes in Computer Science, 15314, 228–243. https://doi.org/10.1007/978-3-031-78341-8_15 https://doi.org/10.1007/978-3-031-78341-8_15 Self-archived version Savic, M., & Zhao, G. (2024). RS-rPPG: Robust self-supervised learning for rPPG. 2024 IEEE 18th International Conference on Automatic Face and Gesture Recognition (FG), 1–10. https://doi.org/10.1109/FG59268.2024.10581991 https://doi.org/10.1109/FG59268.2024.10581991 Self-archived version Savic, M., & Zhao, G. (2025). RS+rPPG: Robust strongly self-supervised learning for rPPG. IEEE Transactions on Circuits and Systems for Video Technology. Advance online publication. https://doi.org/10.1109/TCSVT.2025.3544676 https://doi.org/10.1109/TCSVT.2025.3544676 Self-archived version Savic, M., & Zhao, G. (2023). De-identification of facial videos while preserving remote physiological utility. In British Machine Vision Conference (BMVC), 2023. Advance online publication. https://proceedings.bmvc2023.org/230/ https://proceedings.bmvc2023.org/230/ Self-archived version Savic, M., & Zhao, G. (2024). Oulu remote-photoplethysmography physical domain attacks database (ORPDAD). Lecture Notes in Computer Science, 15131, 51–68. https://doi.org/10.1007/978-3-031-73464-9_4 https://doi.org/10.1007/978-3-031-73464-9_4 Self-archived version Savic, M., & Zhao, G. (2025). Oulu remote-photoplethysmography presentation attacks database (OR-PAD). Manuscript submitted for publication. Tiivistelmä
Etäfotopletysmografia (rPPG) tarjoaa kontaktittoman menetelmän fysiologisten signaalien erottamiseen kasvovideoista, mikä on lupaava vaihtoehto perinteisille kontaktipohjaisille tekniikoille. Tämä opinnäytetyö käsittelee rPPG:n kriittisiä haasteita keskittyen signaalin erottamisen vakauden parantamiseen, yksityisyyden varmistamiseen ja turvallisuuden parantamiseen mahdollisia hyökkäyksiä vastaan.
Vakaan rPPG-mittauksen saavuttamiseksi tämä opinnäytetyö esittelee uusia itseohjautuvia ja transformer-pohjaisia syväoppimismenetelmiä. Erityisesti on kehitetty pitkän aikavälin kontekstin huomioiva transformer-verkko, jota on parannettu monitehtäväisellä esikoulutusstrategialla, mikä osoittaa alan huippusuorituskykyä. Lisäksi on ehdotettu täysin itseohjautuvia kontrastisen oppimisen viitekehyksiä, jotka sisältävät kattavia fysiologisia ennakkotietoja ja datan sisäisiä ja välisiä suhteita saavuttaen suorituskyvyn, joka on kilpailukykyinen ohjattujen menetelmien kanssa useilla eri data-aineistoilla.
Yksityisyys- ja turvallisuushaasteisiin vastaamiseksi tämä opinnäytetyö tutkii sekä tunnistamisen poistamista että hyökkäysalttiutta rPPG-järjestelmissä. Työssä esitellään oppimispohjainen kasvovideon tunnistamisen poistomenetelmä, joka säilyttää rPPG-signaalin eheyden heikentäen merkittävästi biometristä tunnistusta. Lisäksi on luotu kaksi uutta data-aineistoa fyysisen verkkotunnuksen hyökkäysten ja rPPG-pohjaisten esityshyökkäysten kattavaa tutkimista varten, paljastaen kriittisiä haavoittuvuuksia olemassa olevissa rPPG-menetelmissä ja korostaen parannettujen turvatoimien tarvetta.
Yhteenvetona voidaan todeta, että tämä opinnäytetyö edistää merkittävästi rPPG:n alaa kehittämällä vakaita signaalin erotusalgoritmeja, käsittelemällä yksityisyydensuojaan liittyviä huolenaiheita tunnistamisen poistamisen avulla sekä tunnistamalla ja tutkimalla turvallisuusheikkouksia. Nämä edistysaskeleet tasoittavat tietä luotettavammille ja turvallisemmille rPPG-sovelluksille terveydenhuollossa, turvallisuusalalla ja sen ulkopuolella. Osajulkaisut Savic, M., & Zhao, G. (2025). PhySU-Net: Long temporal context transformer for rPPG with self-supervised pre-training. Lecture Notes in Computer Science, 15314, 228–243. https://doi.org/10.1007/978-3-031-78341-8_15 https://doi.org/10.1007/978-3-031-78341-8_15 Rinnakkaistallennettu versio Savic, M., & Zhao, G. (2024). RS-rPPG: Robust self-supervised learning for rPPG. 2024 IEEE 18th International Conference on Automatic Face and Gesture Recognition (FG), 1–10. https://doi.org/10.1109/FG59268.2024.10581991 https://doi.org/10.1109/FG59268.2024.10581991 Rinnakkaistallennettu versio Savic, M., & Zhao, G. (2025). RS+rPPG: Robust strongly self-supervised learning for rPPG. IEEE Transactions on Circuits and Systems for Video Technology. Advance online publication. https://doi.org/10.1109/TCSVT.2025.3544676 https://doi.org/10.1109/TCSVT.2025.3544676 Rinnakkaistallennettu versio Savic, M., & Zhao, G. (2023). De-identification of facial videos while preserving remote physiological utility. In British Machine Vision Conference (BMVC), 2023. Advance online publication. https://proceedings.bmvc2023.org/230/ https://proceedings.bmvc2023.org/230/ Rinnakkaistallennettu versio Savic, M., & Zhao, G. (2024). Oulu remote-photoplethysmography physical domain attacks database (ORPDAD). Lecture Notes in Computer Science, 15131, 51–68. https://doi.org/10.1007/978-3-031-73464-9_4 https://doi.org/10.1007/978-3-031-73464-9_4 Rinnakkaistallennettu versio Savic, M., & Zhao, G. (2025). Oulu remote-photoplethysmography presentation attacks database (OR-PAD). Manuscript submitted for publication. Academic dissertation to be presented with the assent of the Doctoral Programme Committee of Information Technology and Electrical Engineering of the University of Oulu for public defence in the OP auditorium (L10), Linnanmaa, on 22 August 2025, at 12 noonAbstract
Remote photoplethysmography (rPPG) offers a non-contact method for extracting physiological signals from facial videos, presenting a promising alternative to traditional contact-based techniques. This thesis addresses critical challenges encountered in rPPG, focusing on improving signal extraction robustness, ensuring privacy, and enhancing security against potential attacks.
For robust rPPG measurement, this thesis introduces novel self-supervised and transformer-based deep learning methods. Specifically, a long temporal context transformer network is developed, enhanced by a multi-task pre-training strategy, demonstrating state-of-the-art performance. Furthermore, fully self-supervised contrastive learning frameworks are proposed, incorporating comprehensive physiological priors and intra–inter data relationships, achieving performance that is competitive with supervised methods across multiple datasets.
To address privacy and security challenges, this thesis investigates both de-identification and attack vulnerability in rPPG systems. A learning-based facial video de-identification method is introduced, preserving rPPG signal integrity while significantly impairing biometric recognition. Additionally, two novel datasets are created to comprehensively study physical domain attacks and rPPG-based presentation attacks, revealing critical vulnerabilities in existing rPPG methods and highlighting the need for enhanced security measures.
In summary, this thesis contributes significantly to the field of rPPG by developing robust signal extraction algorithms, addressing privacy concerns through de-identification, and identifying and studying security vulnerabilities. These advancements pave the way for more reliable and secure rPPG applications in healthcare, security, and beyond.Tiivistelmä
Etäfotopletysmografia (rPPG) tarjoaa kontaktittoman menetelmän fysiologisten signaalien erottamiseen kasvovideoista, mikä on lupaava vaihtoehto perinteisille kontaktipohjaisille tekniikoille. Tämä opinnäytetyö käsittelee rPPG:n kriittisiä haasteita keskittyen signaalin erottamisen vakauden parantamiseen, yksityisyyden varmistamiseen ja turvallisuuden parantamiseen mahdollisia hyökkäyksiä vastaan.
Vakaan rPPG-mittauksen saavuttamiseksi tämä opinnäytetyö esittelee uusia itseohjautuvia ja transformer-pohjaisia syväoppimismenetelmiä. Erityisesti on kehitetty pitkän aikavälin kontekstin huomioiva transformer-verkko, jota on parannettu monitehtäväisellä esikoulutusstrategialla, mikä osoittaa alan huippusuorituskykyä. Lisäksi on ehdotettu täysin itseohjautuvia kontrastisen oppimisen viitekehyksiä, jotka sisältävät kattavia fysiologisia ennakkotietoja ja datan sisäisiä ja välisiä suhteita saavuttaen suorituskyvyn, joka on kilpailukykyinen ohjattujen menetelmien kanssa useilla eri data-aineistoilla.
Yksityisyys- ja turvallisuushaasteisiin vastaamiseksi tämä opinnäytetyö tutkii sekä tunnistamisen poistamista että hyökkäysalttiutta rPPG-järjestelmissä. Työssä esitellään oppimispohjainen kasvovideon tunnistamisen poistomenetelmä, joka säilyttää rPPG-signaalin eheyden heikentäen merkittävästi biometristä tunnistusta. Lisäksi on luotu kaksi uutta data-aineistoa fyysisen verkkotunnuksen hyökkäysten ja rPPG-pohjaisten esityshyökkäysten kattavaa tutkimista varten, paljastaen kriittisiä haavoittuvuuksia olemassa olevissa rPPG-menetelmissä ja korostaen parannettujen turvatoimien tarvetta.
Yhteenvetona voidaan todeta, että tämä opinnäytetyö edistää merkittävästi rPPG:n alaa kehittämällä vakaita signaalin erotusalgoritmeja, käsittelemällä yksityisyydensuojaan liittyviä huolenaiheita tunnistamisen poistamisen avulla sekä tunnistamalla ja tutkimalla turvallisuusheikkouksia. Nämä edistysaskeleet tasoittavat tietä luotettavammille ja turvallisemmille rPPG-sovelluksille terveydenhuollossa, turvallisuusalalla ja sen ulkopuolella
An Excel-based solution to bring water distribution network analysis closer to users
No full textNowadays it is crucial to align technical research closely with its intended recipients. These recipients are the practitioners who could use new optimal solutions for management purposes in the short term to increase their effectiveness from socio-economic and environmental standpoints. In addition, recipients can be students at any level as in the medium-term they will bring the new management methods and will need to carry out specific technical analyses. The need to transfer and disseminate technological knowledge as soon as it is available is, furthermore, exacerbated by the quick changes that information technology is bringing to the world. Thus, this paper introduces the idea of a collection of MS-Excel add-ins for Water Distribution Network (WDN) analysis. The development of an MS-Excel add-in for WDN analysis makes any technical advance readily usable via a well-known environment. MS-Excel is a friendly environment for users where several add-ins entailing WDN analyses based on classic and recently developed methods can be made available. WDNetXL is a collection of these add-ins. The use of the MS-Excel environment allows the user to personalize their WDNetXL add-in collection and to work easily on input and output of analysed data
Correction: Comparing the temporal dynamics of thematic and taxonomic processing using event-related potentials(PLoS ONE (2017) 12:12 (e0189362) DOI: 10.1371/journal.pone.0189362)
No full textPublisher Copyright: © 2019 Savic et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.Notice of republication An incorrect version of Fig 1 was published in error. This article was republished on June 3, 2019 to correct for this error. Please download this article again to view the correct version
Inside the Rating Scores: A Multilevel Analysis of the Factors Influencing Customer Satisfaction in the Hotel Industry
No full textThese data are used in a research study and may not be redistributed or used for commercial purposes.
Feel free to use it for research purposes or to reproduce the results presented in the article.
When referring to the data set in publications please cite the article as follows:
Radojevic, T., Stanisic, N., & Stanic, N. (2016). Inside the Rating Scores: A Multilevel Analysis of the Factors Influencing Customer Satisfaction in the Hotel Industry. Cornell Hospitality Quarterly.
You may also want to cite the data set:
Radojevic, T., Stanisic, N., & Stanic, N. (2016), Inside the Rating Scores: A Multilevel Analysis of the Factors Influencing Customer Satisfaction in the Hotel Industry., Mendeley Data, v1 http://dx.doi.org/10.17632/kwsrxshf9x.1
For a detailed description of the variables and their descriptive statistics, please read the article
Advances in Data-Driven Analyses and Modelling Using EPR-MOGA
No full textEvolutionary Polynomial Regression (EPR) is a recently developed hybrid regression method that combines the best features of conventional numerical regression techniques with the genetic programming/symbolic regression technique. The original version of EPR works with formulae based on true or pseudo-polynomial expressions using a single-objective genetic algorithm. Therefore, to obtain a set of formulae with a variable number of pseudo-polynomial coefficients, the sequential search is performed in the formulae space. This article presents an improved EPR strategy that uses a multi-objective genetic algorithm instead. We demonstrate that multi-objective approach is a more feasible instrument for data analysis and model selection. Moreover, we show that EPR can also allow for simple uncertainty analysis (since it returns polynomial structures that are linear with respect to the estimated coefficients). The methodology is tested and the results are reported in a case study relating groundwater level predictions to total month-ly rainfall
Gender equality in the context of One Health
No full textMen and women have different roles and society and different access to decision making. In consequence they impact differently their environment and are impacted differently by it. Furthermore, men and women are infected and are affected differently by disease due to their different roles, access to resources and decision-making power. Other factors such as class, skin color, cultural, religious, sexual orientation and linguistic origin inform the way people impact and are impacted by their environment. Understanding how those factors intersect and contribute to different layers of vulnerability is an important aspect if we want to achieve more efficacy in our work and if we want to contribute to a more equitable and inclusive society. Examples related to Obesity, Endocrine Disrupters Chemicals, Anti-Microbial Resistance, COVID 19 and Brucellosis are analyzed
Mile Savic as an interpreter of the recent south Slavic past
No full textThe subject of this paper is the interpretation of the recent South Slavic
past given in the works of Mile Savic, a Serbian philosopher and social
theorist, who recently passed away. The wars for territorial heritage of the
former Yugoslavia, the aggression of the NATO alliance on the Federal
Republic of Yugoslavia, the project of Euro-Atlantic integration of Serbia -
are just some of the most significant thematic points of that interpretation.
By providing an exhaustive analysis of Savic?s attitudes to these and kindred
phenomena of the recent political and social history of the region, the
author concludes that, in a large mosaic of knowledge of a ?time rich in
misfortunes? (Tacitus), the piece attributed to it by this Serbian
philosopher, who left the intellectual and life stage too early, will be, by
all means, among the most significant and precious ones.</jats:p
Afghanistan Under Siege: The Afghan Body and the Postcolonial Border
No full textIn this book, based on field work undertaken in Afghanistan itself and through engagement with postcolonial theory, Bojan Savic critiques western intervention in Afghanistan by showing how its casting of Afghan natives as “dangerous” has created a power network which fractures the country – in echoes of 19th and 20th century colonial powers in the region. Savic also offers an analysis of how and by what means global security priorities have affected Afghan lives
- …
