Research Projects

BRAIN AND SLEEP DISORDERS

Parkinson's Disease

Parkinsonism is a clinical syndrome characterized by bradykinesia, rigidity, tremor, and postural instability at late stage of the disease. (Williams and Litvan, 2013) Broadly, parkinsonian disorders are categorized as neurodegenerative or secondary (when the cause can be recognized such as vascular, infectious, drug induced, etc.). The neurodegenerative parkinsonian disorders are further categorized as idiopathic Parkinson's disease, and atypical parkinsonian disorders such as progressive supranuclear palsy, multiple system atrophy, dementia with Lewy body etc. (Litvan, 2007; Saeed et al., 2020).

Parkinson disease is the second-most common neurodegenerative disorder, impacting 2-3% of the population above 65 years of age. It is considered a long-term gradually progressing neurodegenerative disorder of the central nervous system. It is generally idiopathic, i.e., the cause is unknown. However, genetic forms are also observed. The pathogenesis of Parkinson's disease is not completely known. It is characterized by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta (SNc) which leads to striatal dopamine depletion(Poewe et al., 2017).

The hallmark of Parkinson's disease is characterized by severe progressive loss of neurons from the dopaminergic projections of the substantia nigra pars compacta. Although several structures are involved in Parkinson's disease, the substantia nigra is still considered the most crucial region in PD. We aim to investigate diagnostic and longitudinal biomarkers of PD from early to moderate to advanced stage.

References

  1. Williams and Litvan, 2013
  2. Litvan, 2007
  3. Saeed et al., 2020
  4. Poewe et al., 2017
Isolated Rapid Eye Movement Sleep Behavior Disorders

Rapid Eye Movement (REM) Sleep Behavior Disorders (RBD) is a sleep disorder, first described in 1986, (Schenck et al., 1986) comprising REM sleep parasomnia, featuring unpleasant dreams such as being harmed or hit by someone and/or other strange behaviors such as acting out the dreams in real. RBD is considered prodromal parkinsonism, common in patients with synucleinopathies (60% - 100% of patients) and uncommon in patients with other neurodegenerative disorders. (Arnulf, 2012) Studies have demonstrated more frequent RBD manifestations in males than in females. (Olson et al., 2000; Schenck et al., 1993) Here, we aim to study various aspects leading to conversion of iRBD patients into parkinsonism by identifying multimodal biomarkers at prodromal stage.

Restless Legs Syndrome

Restless legs syndrome (RLS) is a sensorimotor disorder characterized by an urge to move the legs when at rest(1). Although the mechanism of RLS is not well understood, studies have shown a significant dopaminergic system deficit in the basal ganglia of patients with RLS(2). Furthermore, dopamine agonists, which are drugs typically used for Parkinson's disease, provide symptomatic relief.The reported prevalence of RLS in the PD population ranges from 5.5% to 27% in Europe and America(3), which is higher than the 3.9% to 14.3% observed in the background population(4).
The relationship between RLS and PD is still debated. Substantia nigra demonstrates iron increase in PD and iron deficiency in RLS that can be estimated using quantitative susceptibility mapping (QSM) and R2*. Genetics also contributes to the risk of developing RLS and PD. Henceforth, integrating genetics into quantitative and dopamine transporter (DaT) imaging can provide valuable insights. In this project, our aim is to investigate spatiotemporal iron progression and its association with neuromelanin depigmentation, dopaminergic dysregulation and genetic risk factors.

  1. M.M. Ohayon, R. O'Hara, M.V. Vitiello, Epidemiology of restless legs syndrome: A synthesis of the literature, Sleep Med. Rev. 16 (4) (2012) 283-295.
  2. R.M. Rijsman, L.F. Schoolderman, R.S. Rundervoort, M. Louter, Restless legs syndrome in Parkinson's disease, Parkinsonism Relat. Disord. 20 (Suppl. 1) (2014) S5-S9.
  3. D. Verbaan, S.M. van Rooden, J.J. van Hilten, R.M. Rijsman, Prevalence and clinical profile of restless legs syndrome in Parkinson's disease, Mov. Disord. 25 (13) (2010) 2142-2147.
  4. W.G. Ondo, K.D. Vuong, J. Jankovic, Exploring the relationship between Parkinson disease and restless legs syndrome, Arch. Neurol. 59 (3) (2002) 421-424.

DOPAMINERGIC NIGRAL NEUROMELANIN

It was almost a century after Parkinson's disease was first officially portrayed that it was admitted that it is associated with loss of melanized neurons in the substantia nigra.(Tretiakoff C., 1919) Later on, the pigmented neurons were accepted to be dopaminergic showing diminution during Parkinson's disease.(Ehringer and Hornykiewicz, 1960; Poewe et al., 2017)

Interestingly, targeted pigmented noradrenergic neurons in the LC demonstrate similar reduction in Parkinson's disease.(Sulzer and Surmeier, 2013) Hence,From the first official portrayal, it took another century to identify these pigmented neurons in vivo using imaging techniques. Initially, it was proposed that neuromelanin could be a potential marker of SNc damage and MRI-based neuromelanin detection could not only help us in disease diagnosis but also in monitoring the striatal dopamine depletion and thus, can even quantify the nigral damage.(Sulzer et al., 2018; Zecca et al., 2002) This was proposed on the basis of the study on neuromelanin concentration measurement in the SN of HV and PD patients using samples of midbrain taken from deceased people of various age groups, taken during their autopsies within two days of death.(Zecca et al., 2002)

Neuromelanin is an insoluble pigment found in neurons of particular brain regions of various animal species like in humans, appearing around 2-3 years of age and accumulates with aging in humans.(Marsden, 1961; Zecca et al., 2002) The source of neuromelanin is dopaminederived quinones contained within autophagic lysosomes together with lipid bodies and many soluble proteins. Neuromelanin binds high levels of iron and various other metals, forming a paramagnetic complex and hence, becomes visible using neuromelanin-sensitive MRI.(Sasaki et al., 2006; Sulzer et al., 2018) Furthermore, the synthesis of neuromelanin is considered neuroprotective as it removes surplus cytosolic dopamine.(Sulzer et al., 2018) Neuromelanin pigment chelates transition metal ions, including iron, copper, zinc that can help in catecholamine oxidation in the beginning of neuromelanin synthesis and thus, remain bound to neuromelanin in the autolysosomal organelle.(Zecca et al., 1996)

BRAIN IRON

Iron is the most abundant element ferric iron accumulations are not visible in pigmented SN neurons due to effective neuromelanin sequestration.(Zecca et al., 2004a, 2004b) Henceforth, SN is rich in reactive ferric iron but it is challenging to see LC iron which is less rich in overall iron content compared to the SN.(Zecca et al., 2004a) The possible reason why the redox activity of the iron and neurotoxicity are prevented can be explained by this binding to neuromelanin.(Zecca et al., 2008, 1996) Studies have also confirmed the presence of iron inside the neuromelanin organelle using techniques such as electron spectroscopic imaging, nano-secondary ion mass spectrometry and analytical electron microscop.(Biesemeier et al., 2016)

Neuromelanin pigment inside the SN is not detectible under light microscopy in newborns, however it becomes visible at the age of 2-3 years of age,(Fenichel; and Bazelon, 1968) and keeps getting darker through the tenth decade of human lifecycle.(Zecca et al., 2004a, 2002) Although there have been some attempts to study the precise number of neurons having neuromelanin during normal aging using histochemical techniques, results are unclear because sometimes they demonstrated age-dependent diminution and sometimes no variations in the precise number of pigmented neurons.(Cabello et al., 2002; Kubis et al., 2000; Ma et al., 1999) Other chemical methods were also used to compute the concentration of neuromelanin in the SN by separating the pigment and computing ultraviolet absorption of solubilized neuromelanin.(Zecca et al., 2002)
This approach demonstrated an average increase rate of SN neuromelanin concentration during aging around 41 ng/mg SN wet tissue per 98 year.(Zecca et al., 2004a) Normally, the concentration of SN neuromelanin in PD patients is around 50–60% the level of HV possibly because of neuromelanin-based neurodegeneration. Conversely, there is an upsurge in iron concentration in healthy aging in humans, not just in SN but also in putamen, globus pallidus, caudate nucleus, and cortex.(Hallgren and Sourander, 1958; Ramos et al., 2014; Zecca et al., 2004a) alongside maximum iron load in putamen, globus pallidus, and caudate nucleus and minimum in cortical gray matter, white matter, midbrain (including the SN), and cerebellum. However, lowest iron concentrations are found in the pons, medulla and LC.(House et al., 2012; Ramos et al., 2014; Zecca et al., 2004a, 1996)
Such variations related to age also take place in molecular forms of iron, comprising ferritin, transferrin, hemosiderin and neuromelanin and in the distribution of iron compounds between glia and neurons.(Connor et al., 1990; Zecca et al., 2004a, 2001) Ferritin is considered a major iron molecule in most brain cells. Healthy SN aging demonstrates an increase in overall iron load in H-ferritin, L-ferritin, and neuromelanin concentrations whereas in LC, only neuromelanin increases and ferritins and iron are rather lower compared to SN and remains the same during the entire lifecycle.(Zecca et al., 2004a) There is an upsurge in overall iron content in SN in PD patients with iron overload bound to the low-affinity sites of neuromelanin and probable development of a redox active complex(Dexter et al., 1989; Zecca et al., 2008).

Some studies have shown the occurrence of extracellular neuromelanin in the SN in PD patients, particularly demonstrating patterns similar to the extranuclear scattering of the organelles when the neurons were intact.(Surmeier; et al., 2017) During the last years, various iron-sensitive MRI approaches have substantiated the importance of nigral iron increase in PD patients compared to the HV.(Arribarat et al., 2020) Such techniques have even shown overload of iron in both symptomatic and asymptomatic carriers of the LRRK2 and Parkin mutations.(Pyatigorskaya et al., 2015) However, the time to onset before motor sign manifestation is poorly known.
There is a possible close link between dopamine, neuromelanin, and iron concentrations.(Hare and Double, 2016; Zucca et al., 2017) Nonetheless, excess intracellular neuromelanin could also compromise neuronal function and trigger PD-like pathology.(Vila, 2019) Although some studies have reported no correlation between nigral iron and neuromelanin in PD,(Isaias et al., 2016; Reimão et al., 2016) there is a possibility of presence of disease-related iron accumulation in the neuromelanin-based SNc and the natural iron in SNr. Iron deposition in the entire SN, including SNc and SNr can be evaluated using iron-sensitive MRI techniques such as QSM or the apparent transverse relaxation rate (R2*).
These techniques are capable of quantifying SN iron content in PD and iRBD patients, with raised levels often observed when compared with the HV.(Langkammer 99 et al., 2016; Sun et al., 2020) and the natural iron in SNr. Iron deposition in the entire SN, including SNc and SNr can be evaluated using iron-sensitive MRI techniques such as QSM or the apparent transverse relaxation rate (R2*). These techniques are capable of quantifying SN iron content in PD and iRBD patients, with raised levels often observed when compared with the HV.(Langkammer 99 et al., 2016; Sun et al., 2020)

Parkinson's disease demonstrates increased iron concentration in the substantia nigra. The progression of iron and its interaction with neuromelanin content and dopaminergic dysregulation from prodromal to early-stage Parkinson's disease remain poorly understood. Using quantitative susceptibility mapping (QSM) and R2* relaxation rate, we investigated brain iron changes in patients with isolated REM sleep behavior disorder and early-stage Parkinson's disease.

NORADRENERGIC LOCUS

The LC/LsC is a tiny pontine bilateral heterogenous structure in the pons of the brainstem. (Baker et al., 1989; Kaalund et al., 2020) The LC contains melanized noradrenergic neurons, the major source of noradrenaline to the rostral brain. It contributes to arousal, memory and attention, as well as autonomic control. The LsC, located right below the LC, contains neurons driving muscle atonia during REM sleep. Its damage causes RBD. The LC regulates various high order cognition, for example wakefulness/arousal, RBD, pain modulation, local blood movement and immunological tools(Aston-Jones and Bloom, 1981; Benarroch, 2009; Espay et al., 2014; Heneka et al., 2015; O’Donnell et al., 2015), attention, learning, working memory (Robbins, 1984; Aston-Jones and Cohen, 2005; Benarroch, 2009; Mather et al., 2016). Also, age-related decline within the LC–based noradrenergic system is linked to reduction in cognitive capacities pertaining to episodic memory (Ha¨mmerer et al., 2018, Jacobs et al., 2018b; Dahl et al., 2019).

The LC/LsC variations in different neurodegenerative diseases suggest that it may be a susceptible target for neurodegenerative pathophysiology (Sharma et al., 2010). Nonetheless, more studies are warranted to understand the reason. This could be either due to high metabolic need of neurons in the LC/LsC which is necessary to maintain fundamental physiological functions or due to the vicinity of the LC/LsC to the fourth ventricle, revealing it to toxins from the CSF (Mather and Harley, 2016; Weinshenker, 2018). Although the LC/LsC is considered one of the first brain structures to display neurodegenerative disease pathology, little is known about how the structural and functional changes in the LC/LsC influence the course of pathogenesis and symptoms. However, it has been hypothesized that hyperactivity in damaged LC/LsC might speed up the neuropathological propagation in neurodegenerative diseases via noradrenergic projections (Weinshenker, 2018). Timely identification of the LC/LsC damage using imaging techniques could be useful in early diagnosis of the disease and help in pharmacological interventions to improve hyperactivity of the noradrenaline system and possibly reduce the speed of progression of the disease.

Lately, it has been demonstrated that the LC/LsC variations can be studied using neuroimaging techniques using MRI. (Sasaki et al., 2006 ; Keren et al., 2009; Clewett et al., 2016; Betts et al., 2017; Priovoulos et al., 2018). The LC/LsC imaging can act as a future biomarker for noradrenergic dysfunction and could not only be useful for better staging of neurodegenerative disorders but also in clinical trials targeting the noradrenergic system for predicting treatment success. (Betts et al., 2019) 104 Interestingly, there have been several efforts to establish LC MRI as a powerful biomarker for diagnosis and targeting therapeutic interventions in neurodegeneration, particularly during the first European LC Imaging meeting held in Magdeburg in 2018, the challenges associated with introducing the LC MRI as a biomarker was deeply discussed. (Betts et al., 2019) The LC/LsC can be seen using MRI that are sensitive to neuromelanin that stores noradrenergic neurons and are a sign of LC/LsC integrity, that is the cell density). (Betts et al., 2019) Several studies have demonstrated that such neuromelanin-based MRI technique can distinguish patients from the HV with the help of LC/LsC-based signal differences, particularly in PD, Alzheimer's disease, major depression (Liu et al., 2017).

NIGROSOME

Nigrosome 1 Early diagnosis of differential diagnosis in parkinsonism-related disorders such as dementia with Lewy bodies (DLB) from non-parkinsonianconditions such as Alzheimer’s disease (AD), is essential for effective clinical management . Studies have demonstrated that substantia nigra parscompacta (SNc) can be divided into a matrix and five clusters with high neuromelanin concentration, namely nigrosomes (1-5) . In Parkinson'sdisease (PD), the nigrosome 1 (N1) plays a crucial role as it gets affected first. In healthy individuals, N1 partially overlaps with the hyperintensedorsolateral ‘swallow tail sign’ (STS) on iron-sensitive MRI . The absence of this sign achieves over 90% diagnostic accuracy for distinguishing PDand related disorders from non-parkinsonian ones . SPECT/PET-DOPA imaging remains the gold standard but is costly and often inaccessible andSTS disappearance may be an early indicator of neurodegeneration. Despite its potential, the STS remains underutilized, mainly due to the need forspecialized rater training and the lack of reliable segmentation methods . AI techniques have demonstrated robust results in segmenting SNcusing neuromelanin imaging

We present NigrosomeNet, a fully automated tool for STS segmentation designed to enable fast, accurate differentiation of neurodegenerativediseases.

  1. Poewe W, Seppi K, Tanner CM, et al. Parkinson disease. Nat Rev Dis Prim 2017;3:1-21
  2. Damier P, Hirsch EC, Agid Y, Graybiel AM. The substantia nigra of the human brain. II. Patterns of loss of dopamine-containing neurons in Parkinson's disease. Brain 1999;122:1437-1448
  3. A.I. Blazejewska, S.T. Schwarz, A. Pitiot, M.C. Stephenson, J. Lowe, N. Bajaj, R. W. Bowtell, D.P. Auer, Visualization of nigrosome 1 and its loss in PD:pathoanatomical correlation and in vivo 7 T MRI, Neurology 81 (6) (2013) 534-540.
  4. Schwarz ST, Afzal M, Morgan PS et-al. The 'swallow tail' appearance of the healthy nigrosome - a new accurate test of Parkinson's disease: a case-control and retrospective cross-sectional MRI study at 3T. PLoS ONE. 2014;9 (4): e93814. doi:10.1371/journal.pone.0093814 -
  5. Brammerloh M, Kirilina E, Alkemade A, et al. Swallow tail sign: Revisited. Radiology 2022;305(3):674-677.
  6. Shams S, Fällmar D, Schwarz S, Wahlund L-O, van Westen D, Hansson O, Larsson E-M, Haller S. MRI of the Swallow Tail Sign: A Useful Marker inthe Diagnosis of Lewy Body Dementia?. American Journal of Neuroradiology. 38 (9): 1737. doi:10.3174/ajnr.A5274
  7. Cheng Z, Zhang J, He N, Li Y, Wen Y, Xu H, Tang R, Jin Z, Haacke EM, Yan F, Qian D. Radiomic Features of the Nigrosome-1 Region of the SubstantiaNigra: Using Quantitative Susceptibility Mapping to Assist the Diagnosis of Idiopathic Parkinson's Disease. (2019) Frontiers in aging neuroscience. 11:167. doi:10.3389/fnagi.2019.00167
  8. Kang, J., Kim, H., Kim, Eunjin, Kim, Eunbi, Lee, H., Shin, N., Nam, Y., 2021. Convolutional Neural Network-Based Automatic Segmentation ofSubstantia Nigra on Nigrosome and Neuromelanin Sensitive MR Images. Investig. Magn. Reson. Imaging 25, 156.
  9. Pastor, M.A., Ortiz de Solórzano, C., 2024. AutomaticSegmentation and Quantification of Nigrosome-1 Neuromelanin and Iron in MRI: A Candidate Biomarker for Parkinson's Disease. J. Magn. Reson.Imaging 60, 534-547.
  10. Gaurav, R., Valabrègue, R., Yahia-Chérif, L., Mangone, G., Narayanan, S., Arnulf, I., Vidailhet, M., Corvol, J.-C., Lehéricy, S., 2022. NigraNet: AnAutomatic Framework to Assess Nigral Neuromelanin Content in Early Parkinson's Disease Using Convolutional Neural Network. NeuroImage Clin.36, 103250.
  11. Shinde, S., Prasad, S., Saboo, Y., Kaushick, R., Saini, J., Pal, P.K., Ingalhalikar, M., 2019. Predictive markers for Parkinson's disease using deepneural nets on neuromelanin sensitive MRI. NeuroImage Clin. 22, 101748
  12. Ronneberger, O., Fischer, P., Brox, T., 2015. U-net: Convolutional networks for biomedical image segmentation. Lect. Notes Comput. Sci.(including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics) 9351, 234-241.
  13. Perez-García, F., Sparks, R., Ourselin, S., 2020. TorchIO: a Python library for efficient loading, preprocessing, augmentation and patch-basedsampling of medical images in deep learning.

DEEP LEARNING IN CLINICAL NEUROSCIENCE

We have interdisciplinary interest and considering my background with a judicious mix of engineering methods and clinical neuroscience, we leverage these ideas to develop most useful, accurate and reliable AI-based biomarkers
Currently, deep learning (DL) is used in a wide array of image processing tasks from image segmentation to extraction of various kinds of biomarkers, such as signal intensity and volume of brain structures or lesions. (Burgos and Colliot, 2020; Goodfellow, I., Bengio, Y., Courville, A., & Bengio, 2016; Lecun et al., 2015) Furthermore, in DL, convolutional neural networks (ConvNet) use a simple artificial neural network architecture that has been successfully applied to image segmentation, recognition and detection of region of interest (ROI)s in medical images.(Cun; et al., 1998; Krizhevsky et al., 2012)

With respect to manual methods, ConvNet approaches can improve consistency, be less prone to errors and save a lot of time. While a few studies have shown the usefulness of ConvNet-based automatic SNc ROI segmentation approaches, they were tested on rather small datasets of PD patients by training the model using data augmentation techniques. (Krupička et al., 2019; Le Berre et al., 2019; Shinde et al., 2019). Therefore, there is a need for a robust automatic SNc segmentation framework tested on a larger dataset of PD patients, involving fewer training data and no data augmentation that can still yield diagnostic performance similar to established ground truth, manual SNc segmentation performed by experts.

In this study, we propose a framework, NigraNet, that facilitates fully automatic SNc segmentation by employing a ConvNet-based architecture based on a modified U-net.(Ronneberger et al., 2015) We aimed to investigate the neuromelanin changes in early-stage idiopathic PD patients as compared to HV, as a biomarker of SNc neurodegeneration. Thereafter, we also compared the automatic segmentation measurements to the manual measurements for the validation of our proposed framework.

AGING AND GLOBAL HEALTH

According to the World Population Ageing report of United Nations, the population above the age of 60 years will rise to over 2 billion by 2050 though it was 841 million in 2013. (United, 2019) This report also emphasizes that 1 in 6 people in the world will be over the age of 65 by 2050 that used to be 1 in 11 in 2019. Interestingly, the older population of our planet is not only growing in absolute terms but also in relative terms. Every corner of the world will see an increase in the size of their older population between 2019 and 2050 with the largest increase of +312 million people projected in Eastern and South-Eastern Asia, rising from around 261 million in 2019 to 573 million people aged 65 years or more in 2050 (Table 1). (United, 2019)

It can be deduced from these population growth figures projected in Table 1 that this alarming rate of increase in age will lead to a sharp increase in many neurodegenerative diseases such as Parkinson's disease (PD), Alzheimer's disease etc. As demonstrated in Figure 1, from 1990 to 2015, the number of people with PD doubled to over 6 million.(Ray Dorsey et al., 2018) Driven principally by aging, this number is projected to double again, to over 12 million by 2040. Thus, as populations age at this rate, it will be critical to ensure access not only to disease prevention but also to screening and treatment as needed. Healthy aging without neurodegenerative diseases could become a more difficult mission in the future. Henceforth, this leads us to a vital need to develop reliable and accurate biomarkers, particularly dedicated to early disease detection in order to help us understand the pathophysiology of such neurodegenerative diseases and also to facilitate the application of timely symptomatic disease-modifying interventions.(Betts et al., 2019)

This is one of the most ambitious ongoing projects. The aim is to perform human brain mapping trajectories across the adult lifespan and also try to understand whether these patterns are ancestry-specific particularly focusing on underrepresented populations.

PAST PROJECTS

SELECTIVE ATTENTION

This was a project at the département d'architecture, de design et de technologie des médias. The research project was focused in human attention using the subjective experience of players in a video game.
expérience de jeu subjective : combattre l'attention sélective avec des débriefings vidéo à l'aide de dispositifs de suivi oculaire et de surveillance de la fréquence cardiaque. expériences psychophysiques impliquant des sujets humains pour obtenir des mesures de qualité subjectives et objectives en utilisant de jeu vidéo.

SELECTIVE ATTENTION

This was a project at the département d'architecture, de design et de technologie des 5 / 2013 – 06 / 2013 : Stagiaire en recherche Université Dunarea de Jos de Galati, Roumanie
Unité de recherche : Département de chimie, physique et environnement Direction Pr. Luminita Moraru Il regroupe une trentaine de personnes.
Thématiques : Développement de nouvelles méthodes de calcul, incluant la reconstruction / analyse d'images médicales, les méthodes informatiques en imagerie médicale, le traitement et la segmentation d'images médicales, la détection / diagnostic assisté par ordinateur, les techniques de construction d'images médicales et l'imagerie en radiologie / échographie diagnostique.
Mes Missions : Exécuter des projets de recherche Le projet de recherche sur la segmentation de l'image IRM cérébrale.
Mes Activités : Ø Élaborer des modèles théoriques sur le projet.
Ø Préparer la réalisation pratique
Ø Acquisition des données d’IRM
Ø Gestion des données acquises

  • Stockage et transfert de la base de données.
  • Effectuer l'analyse des données et contrôle de la qualité des données.
  • Programmation avec MATLAB pour segmentation.
  • Mise en commun des données
  • Ø Valorisation et diffusion des résultats
  • Synthétiser, mettre en forme et valider les résultats.
  • Rédaction d’articles originaux sur neuroimagerie.

02 / 2011- 02 / 2013: Assistant de Recherche ISIK Université, Istanbul, Turquie
Unité de recherche: Conseil de la recherche scientifique et technologique de Turquie (TÜBİTAK) et Département de génie électrique et électronique Direction L’équipe d’ISIK Université était dirigée par Pr. Ahmet Aksen. 24
Thématiques : Les domaines de recherche du département incluaient le génie électrique, électronique et biomédical, en se concentrant particulièrement sur le traitement de l'image et de la vidéo. TÜBİTAK est une agence nationale turque dirigée par un conseil scientifique axées sur la vision par ordinateur et l'apprentissage automatique, l'apprentissage en profondeur, le traitement d'image / vidéo, l'imagerie médicale et plus.
Mes Missions : Exécuter des projets de recherche. Le projet de recherche était axé sur l’amélioration de la résolution vidéo assistée par encodeur pour les vidéos compressées et évaluation de la qualité à l'aide de mesures subjectives et objectives
Mes Activités : Ø Élaborer des modèles théoriques

  • Formuler des hypothèses de recherche et réaliser des revues de la littérature scientifique sur le traitement d’image et vidéo.
  • Initiation au travail de bibliographie, de revue de la littérature scientifique.
    • Ø Préparer la réalisation pratique
  • Étude de la faisabilité pratique (répartition des coûts, organisation des horaires, définition des sites d'action pour l’expérimentation).
  • Coordination pratique du projet entre les différentes parties prenantes.
  • Rôle de recrutement des volontaires sains avec entretien de recrutement.
  • Préparer les produits techniques et les supports.
    • Ø Acquisition des données
  • Mener des expériences psychophysiques impliquant des sujets humains pour obtenir des mesures de qualité subjectives et objectives.
    • Ø Gestion des données acquises
  • Stockage et transfert de la base de données.
  • Effectuer l'analyse des données et contrôle de la qualité des données.
  • Surveiller les résultats intermédiaires et adapter le protocole de recueil des données.
  • Assurer la traçabilité et la reproductibilité des résultats.
  • Programmation avec MATLAB et C++ (OpenCV).
  • Mise en commun des données
    • Ø Valorisation et diffusion des résultats
  • Synthétiser, mettre en forme et valider les résultats.
  • Rédaction des documents internes et les papiers de conférences.
  • Travailler en coordination scientifique.
  • Rédaction de mémoire de Masters.
    • MEDUSA

XLcloud Unité de recherche : ARTEMIS - Advanced Research and TEchniques for Multidimensional Imaging Systems Direction Pr. Titus Zaharia Il regroupe aujourd'hui une trentaine de personnes.

Thématiques : Le coeur des recherches d'ARTEMIS relève des sciences et technologies de l'image numérique. ARTEMIS traite de la chaîne de l'image depuis la création des contenus numériques jusqu'à leur diffusion. L'enjeu est de créer, modéliser, analyser, indexer, animer, sécuriser, manipuler, enrichir, coder, distribuer et visualiser des contenus hétérogènes et complexes pour des services d'intermédiation économiquement réalistes.
Mes Missions : Exécuter et mener des projets de recherche Le projet XLcloud financé par le FSN français (Fonds national pour la société numérique) programme visait à définir et à démontrer les principes du calcul haute performance en tant que service pour toutes les applications qui impliquent des calculs très intensifs, le rendu 3D, le streaming vidéo et la visualisation à distance.
Mes Activités : Ø Élaborer des modèles théoriques

  • Formuler des hypothèses de recherche et comparer des modèles théoriques aux données de la littérature et réalisation de revues de la littérature scientifique.
  • Formation à la lecture critique d'articles scientifiques spécifiquement dédiés au traitement d'imagerie et vidéo.
  • Initiation au travail de bibliographie, de revue de la littérature scientifique
    • Ø Préparer la réalisation pratique de l'étude
  • Étude de la faisabilité pratique.
  • Participation à des réunions de recherche hebdomadaires.
    • Ø Acquisition des données
  • Mener des expériences psychophysiques impliquant des sujets humains pour obtenir des mesures de qualité subjectives et objectives.
    • Ø Gestion des données acquises 22
  • Stockage et transfert de la base de données.
  • Effectuer l’analyse des données et mise en oeuvre d'applications sur la plate-forme commune d’ARTMEIS.
  • Contrôle de la qualité des données.
  • Programmation avec MATLAB.
  • Mise en commun des données
    • Ø Valorisation et diffusion des résultats
  • Synthétiser, mettre en forme et valider les résultats.
  • Communication régulière avec les partenaires du projet XLcloud.
  • Rédaction des documents internes et de livrables.