Vous êtes ici : Accueil » Kiosque » Annonce

Identification

Identifiant: 
Mot de passe : 

Mot de passe oublié ?
Détails d'identification oubliés ?

Annonce

5 mars 2019

Addressing Challenges of Machine Learning in Medical Image Analysis


Catégorie : Post-doctorant


The purpose of the post-doctoral position is to propose innovative solutions to address the particular challenges that deep learning has when confronted to medical image data, in particular regarding the collection of annotations, dealing with small amounts of annotated data, and the need for certainty measurements.

The position will be hosted by the LS2N Lab in Nantes (site Centrale Nantes) for 12 months (renewable 12 months)

 

 

 

 

Addressing Challenges of Machine Learning in Medical Image Analysis

Postdoctoral Position

 

Current trends in medical image analysis have shown the effectiveness of Machine Learning in devising computer-aided solutions for a plethora of applications and imaging modalities. However, the success of ML approaches depends today upon the availability of large volumes of data and moreover of expert annotations. Collecting such expert annotations is a limiting bottleneck for many applications, where it is unfeasible to gather a significant number of annotated examples of good quality, and considering there is never a full consensus among experts. Even when data and annotations are available, transferability of successful solutions to clinics remains an issue. In particular, the adoption of deep learning approaches has been slowed down by the lack of certainty measures guaranteeing the safety of the predicted decisions. Finally, there are all the ethical issues regarding the collection, transfer and analysis of

 

The purpose of the post-doctoral position is to propose innovative solutions to address the particular challenges that deep learning has when confronted to medical image data, in particular regarding the collection of annotations, dealing with small amounts of annotated data, and the need for certainty measurements.

 

The post-doc is open to propositions heading to one among the following directions:

-Continuous, interactive and active learning methods, for instance involving both algorithms and physicians into the learning loop.

-Uncertainty modelling of annotations and learning models.

-Exploiting weaker levels of annotations: transfer-learning, semi-supervised learning, etc.

 

The methods proposed during the research project should either:

-demonstrate a reduction of the annotation time,

-reduce the need for large annotated datasets,

-provide uncertainty measures along predictions,

while maintaining good accuracy and ensuring usability.

 

The successful candidate will work in close collaboration with medical partners from the CHU hospital in Nantes towards CAD solutions in nuclear medicine. Other. She/He will integrate the recently granted MILCOM project with at least 4 Ph.D. students and two post-docs. She/He will have the opportunity to work hand in hand with graduate students and undergraduate students. There is also the potential for establishing collaborations with industrial and international partners.

 

The position will be hosted by the laboratory of digital sciences of Nantes ("Laboratoire des Sciences du Numérique de Nantes: LS2N" in French) at the Central Nantes location. The candidate will integrate the SIMS(www.ls2n.fr/equipe/sims), a research team. SIMS focuses on the conception of methodological tools and the exploitation of the intrinsic structure of data through statistical processing of signal and images.

 

"Regularly quoted in newspapers as being one of the nicest cities in France, Nantes is also renowned for being a rich, lively and innovative city. Its economic clout makes Nantes France's 3rd largest industrial city and 2nd most successful city in terms of employment growth."(https://en.nantes.fr/home.html)

 

Requirements for application are

-a solid background in medical image analysis.

-experience in development of deep learning algorithms.

 

The position is for 12 months (renewable +12 months).

Starting date: open from April 2019 but flexible.

 

Contact: Diana Mateusdiana.mateus@ec-nantes.fr

 

Dans cette rubrique

(c) GdR 720 ISIS - CNRS - 2011-2019.