• Aucun résultat trouvé

PhD student in Deep Learning applied to polarimetric imaging for object detection in road scenes

N/A
N/A
Protected

Academic year: 2022

Partager "PhD student in Deep Learning applied to polarimetric imaging for object detection in road scenes"

Copied!
2
0
0

Texte intégral

(1)

RACHEL BLIN

PhD student in Deep Learning applied to polarimetric imaging for object detection in road scenes

[ rachel.blin@insa-rouen.fr R 29 rue aux Ours, 76000 Rouen, France ¯ linkedin.com/in/rachel-blin-2b0a07149 Ó +33 6 58 48 01 53 ‡ github.com/RachelBlin ® http://pagesperso.litislab.fr/rblin/

PROJECTS

INSA Certified ISO 9001:2015 Project January 2017 - January 2018

• Working for the client A2iA, development of handwriting recognition software company

• Creating an engine to process a PDF document to return its written content

• Quality manager of the project

Image processing

September 2017 - January 2018

• Extraction of the reflection of an image using the KNMF algorithm

September 2016 - January 2017

• Image segmentation depending on its different textures

TECHNICAL SKILLS

Programming languagesPython, Matlab, Java, SQL, C, Pascal, Bash

FrameworksKeras, NumPy, SciPy, Pandas

AlgorithmsDeep Neural Networks for object detection, Decision trees, SVM, K-NN

DatabasesMySQL, PostgreSQL

Web technologiesHTML5, CSS3, Computer networking

LANGUAGES

• French : native speaker

• English : Full Professional Proficiency (TOEIC 905/990)

• Spanish : Professional Working Proficiency

• Arabic : Limited Working Proficiency

EXPERIENCE ABROAD

Africa: Cameroon (1996-2001,

2011-2012), Nigeria (2006-2009), Libya (2009-2010), Senegal (2020)

North America: USA (2017)

Asia: Syria (2001-2003), Lebanon (2010-2011)

Europe: Spain (2005-2006)

EDUCATION

PhD in Deep Learning applied to polarimetric imaging for object detection in road scenes

LITIS lab, INSA Rouen Normandie

since September 2018 ½ Rouen, France Co-advised by Dr. Samia Ainouz and Prof. Stéphane Canu

Engineering school in data science

INSA Rouen Normandie, Information Systems Architecture department

September 2013 - July 2018 ½ Rouen, France

First year of medicine studies

Université Pierre et marie Curie

September 2012 - June 2013 ½ Paris, France

PROFESSIONAL EXPERIENCE

Researcher in Deep Learning applied to polarimetric imaging

LITIS lab, INSA Rouen Normandie

since September 2018 ½ Rouen, France

• Knowledge of the deep neural networks used for object detection and train some of them on my own datasets

• Introduction to polarimetric imaging and extraction of the most relevant features to discriminate an object

• Creation of the first public large polarimetric dataset labelled for object detection purpose and containing diverse weather conditions

• Teaching

• Supervising an intern

Engineer internship as junior architect

Sopra HR Software

Since February 2018 ½ La Défense, France

• Design of new solutions for payroll control

• Exploration of technologies for data analysis and machine learning

• Test the prototype to solve the problem

Specialization internship as web designer

New York University

June 2017 - August 2017 ½ New York, USA

• Design of a website for the Diet Optimizer project, a linear program to give a user recipes to meet his nutritional requirements

• Creation of MySQL databases to make a profile for the user allowing him to give his feed-back on recipes

• Use of data science optimization and research oriented approach

(2)

PUBLICATIONS

A new multimodal RGB and polarimetric image dataset for road scenes analysis

Rachel Blin, Samia Ainouz, Stéphane Canu, Fabrice Meriaudeau 2020 Conference on Computer Vision and Pattern Recognition Workshop (CVPR)

Road scenes analysis in adverse weather conditions by polarization-encoded images and adapted deep learning Rachel Blin, Samia Ainouz, Stéphane Canu, Fabrice Meriaudeau 2019 IEEE Intelligent Transportation Systems Conference (ITSC)

Adapted learning for polarization-based car detection

Rachel Blin, Samia Ainouz, Stéphane Canu, Fabrice Meriaudeau Fourteenth International Conference on Quality Control by Artificial Vision (QCAV), 2019

TEACHING

Numerical Analysis

• 2019-2020 school year at INSA Rouen Normandie, practical sessions

• February 2020 at Virtual University of Senegal, lecture courses and practical sessions

Information Theory

• 2019-2020 school year at INSA Rouen Normandie, practical sessions

• Design of the practical sessions

Signal processing

• 2019-2020 school year at INSA Rouen Normandie, practical sessions

Introduction to Deep Learning

• 2019-2020 school year at INSA Rouen Normandie, lecture courses and practical sessions

Références

Documents relatifs

From left to right : left image, ground truth disparity map, edge image, disparity map computed based on cross based adaptive region, and disparity map obtained via the modified

Pour le comprendre, il nous faut d’abord évoquer les finesses sémantiques de ce domaine, depuis l’intelligence artificielle vers l’apprentissage automatique * (machine learning)

In order to compare the detection in road scenes in adverse weather conditions using RGB and polarimetric encoded images, we carried the following experiment.. Before going into

For each category of learning setup, i.e., supervised, semi-supervised, and unsu- pervised, one ML-based and one DL-based technique are evaluated and compared in terms

The 3D semantic mapping is described in Section 4 , including the semantic segmentation, the monocular visual SLAM, the Bayesian incremental fusion, and the global

Here, our aim is to first investigate the structure of weighted niche-overlap graph using four classical metrics: node degree, clustering coefficient, assortativity and

La note d’état corporel (NEC) est, avec le poids, le meilleur indicateur pratique pour estimer la part de masse grasse de l’animal. L’accès à cette masse grasse permet

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des