Asma Rejeb Sfar

Title: Asma Rejeb Sfar: Innovator in Machine Learning and 3D Object Detection

Introduction

Asma Rejeb Sfar is a prominent inventor based in Paris, France. She has made significant contributions to the fields of machine learning and computer vision, holding a total of nine patents. Her innovative work focuses on improving the detection and segmentation of 3D objects and 2D floor plans.

Latest Patents

One of her latest patents is a method for machine-learning for 3D object detection. This computer-implemented method involves learning a neural network that encodes a super-point of a 3D point cloud into a latent vector. The process includes obtaining a dataset of super-points, where each super-point represents a part of an object. The learning process minimizes a loss that penalizes disparities between two super-points, leading to improved machine-learning capabilities for 3D object detection.

Another notable patent is related to semantic segmentation of 2D floor plans using a pixel-wise classifier. This method determines a function for semantic segmentation of a 2D floor plan that represents a building's layout. It involves providing a dataset of 2D floor plans, each associated with a respective semantic segmentation, and learning the function based on this dataset. This approach offers an enhanced solution for processing 2D floor plans.

Career Highlights

Asma has worked with renowned companies such as Dassault Systèmes, where she has applied her expertise in machine learning and computer vision. Her work has significantly impacted the development of innovative technologies in these fields.

Collaborations

Asma has collaborated with talented individuals, including Malika Boulkenafed and Louis Dupont De Dinechin. These collaborations have further enriched her research and development efforts.

Conclusion

Asma Rejeb Sfar is a trailblazer in the realm of machine learning and 3D object detection. Her patents and career achievements reflect her dedication to advancing technology and innovation.