Hi, I am Rachid

Rachid Zeghlache

Passionate about new technologies to enhance medicine, especially medical imaging. Holder of an engineering degree in computer science and Technologies for Health and a Master 2 in Images and Signals for Medicine. Currently looking for a PhD position in medical imaging and AI. I have already worked in multidisciplinary research environments on common issues related to Data science and healthcare, like segmentation of organs at risks or classification/detections of cancerous lesions.

Leadership
Team Work
Communication
Hard Working
Fast Learner
Problem Solving

Skills

Experiences

1
Machine Learning Engineer
Laboratoire Images, Signaux et Systèmes Intelligents (LISSI)

October 2020 - December 2020, Vitry sur Seine, France

LISSI develops multidisciplinary research, both theoretical and applied, in the field of information and communication sciences and technologies and in particular artificial intelligence. The applications dealt with are mainly in the field of technologies for health and well-being and are focused on difficult issues such as assistance with diagnosis and therapeutic monitoring, ageing, assistance to dependent or disabled people and e-health.

Responsibilities:
  • Scientific and technology watch on state-of-the-art models used in mental state detection with EEG data.
  • Application of classical classification machine learning algorithms.
  • Development of deep learning models for features extraction and mentale state classification.
Intern Data scientist
Nuclear medicine departmart of the Henri Mondor Hospital.

March 2020 - September 2020, Créteil, France

Nuclear medecine department which use deep learning and bayesian model to help doctors in their clinical routine.

Responsibilities:
  • Scientific and technology watch on state-of-the-art generatives models and object detections models used in medical imaging.
  • re-implementation in pytorch of the Hierarchical probabilistic U-net in both 2D and 3D.
  • Fusion of the HPU-net in 2D with the DeTr.
2
3
Intern Data scientist
German Cancer Research Center (DKFZ).

Sep 2019 - Jun 2019, Heidelberg, Germany

Intern in computer vision and deep learning for segmentating bones.

Responsibilities:
  • Scientific and technology watch on state-of-the-art segmentation methods for medical imaging.
  • Development of two deep learning models for segmentation (U-net) one with pixels wise (2D slices) and the second with voxels wise (3D patchs).
  • Analyse of the influence of data augmentation technics, classic transformation Vs Elastique one.

Projects

KnowUrBones
Developer June 2019 - September 2019

Developpment of program that segment head and neck bones on CT scans based on 2D U-net and 3D U-net.

Star
Master's these project
Developer March 2020 - September 2020

Segmentation of liver in SPECT images using deep learning.

Star
Master graduation project
Developer Jan 2020 - Feb 2020

Detection and segmentation of lesions in PET scan using deep bayesian segmentation models and one stage object detector.

Star
Detection of mental state while driving using EEG data, machine learning and bayesian deep learning.
Owner Oct 2020 - Dec 2020

Recent Posts

Card image cap
R-CNN Family

This sample post tests the followings: Category, sub-category nesting in the sidebar. Hero image and other images are in images folder inside this post directory. Different media rendering like image, tweet, YouTube video, Vimeo video etc. Image Sample Tweet Sample “In addition to being more logical, asymmetry has the advantage that its complete appearance is far more optically effective than symmetry.” — Jan Tschichold pic.twitter.com/gcv7SrhvJb — Graphic Design History (@DesignReviewed) January 17, 2019 YouTube Video Sample Vimeo Video Sample

September 8, 2020 Read
Card image cap
Object detection in deep learning

The goal is to predict the localization of objects in an image via bounding boxes and the classes of the located objects. With deep learning approach we tackle this problem with two type of models called one stage or two stages detectors. Even if they share a common gold the two approach are sightly different and can be summarized as follow: Common tricks used in object detection model Bounding Box Regression We define by p $=(p_{x},p_{y},p_{w},p_{h})$ the prediction of the bounding box regressor, where x and y correspond to the centred coordinate and h and w its height, weight the corresponding ground truth box coordinates g $=(g_{x},g_{y},g_{w},g_{h})$, the regressor is configured to learn scale-invariant transformation between two centers and log-scale transformation between widths and heights.

September 8, 2020 Read
Card image cap
Object detection in deep learning

The goal is to predict the localization of objects in an image via bounding boxes and the classes of the located objects. With deep learning approach we tackle this problem with two type of models called one stage or two stages detectors. Even if they share a common gold the two approach are sightly different and can be summarized as follow: Common tricks used in object detection model Bounding Box Regression We define by p $=(p_{x},p_{y},p_{w},p_{h})$ the prediction of the bounding box regressor, where x and y correspond to the centred coordinate and h and w its height, weight the corresponding ground truth box coordinates g $=(g_{x},g_{y},g_{w},g_{h})$, the regressor is configured to learn scale-invariant transformation between two centers and log-scale transformation between widths and heights.

September 8, 2020 Read

Achievements

Hackathon E-Med, Erreurs médicamenteuses ANSM

Datathon Janssen, défi Oncologie Johnson Johnson