Assia Benbihi

Postdoc at the Spatial Intelligence Group

Bio

I am interested in 3D computer vision and how to develop methods that work in the real world. Some of the challenges I work on include large-scale and object-scale 3D reconstruction, long-term localisation and mapping, localisation in natural environments and across extreme appearance variations, and privacy-preserving localisation. Feel free to reach out if you want to chat about these topics.

Currently, I am a postdoc at the Spatial Intelligence Group led by Torsten Sattler.. Prior to that, I interned at Niantic, worked at the Visual Recognition Group, and received my PhD from CentraleSupelec at the DREAM Lab, under the supervision of Cédric Pradalier and Matthieu Geist, tackling long-term localisation problems across seasons in challenging outdoor environments.

Publications

- EdgeGaussians -- 3D Edge Mapping via Gaussian Splatting
- Kunal Chelani, Assia Benbihi, Torsten Sattler, Fredrik Kahl
- Preprint 2024
- Obfuscation Based Privacy Preserving Representations are Recoverable Using Neighborhood Information
- Kunal Chelani^o, Assia Benbihi^o, Fredrik Kahl, Torsten Sattler, Zuzana Kukelova
- Preprint 2024
- Comparative Evaluation of 3D Reconstruction Methods for Object Pose Estimation
- Varun Burde^o, Assia Benbihi^o, Pavel Burget, Torsten Sattler
- Preprint 2024
- Differentiable Product Quantization for Memory Efficient Camera Relocalization
- Zakaria Laskar, Iaroslav Melekhov, Assia Benbihi, Shuzhe Wang, Juho Kannala
- ECCV 2024
- Gaussian Latent Representations for Uncertainty Estimation using Mahalanobis Distance in Deep Classifiers
- Aishwarya Venkataramanan, Assia Benbihi, Martin Laviale, Cédric Pradalier
- ICCVW 2023
- Object-Guided Day-Night Visual Localization in Urban Scenes
- Assia Benbihi, Cédric Pradalier, Ondrej Chum
- ICPR 2022
- Robust Visual Features for Long-Term Monitoring (PhD Thesis)
- Assia Benbihi
- 2020
- WASABI: Image-Based Place Recognition on Bucolic Environment Across Seasons From Semantic Edge Description
- Assia Benbihi, Stephanie Arravechia, Matthieu Geist, Cédric Pradalier
- International Conference on Robotics and Automation 2020
- Code, Video
- ELF: Embedded Localisation of Features in pre-trained CNN
- Assia Benbihi, Matthieu Geist, Cédric Pradalier
- International Conference on Computer Vision 2019
- Code, Video
- Semantic Nearest Neighbor Fields for Monocular Edge Visual-Odometry
- Xiaolong Wu, Assia Benbihi, Matthieu Geist, Cédric Pradalier
- Preprint 2019
- Semi-Supervised Domain Adaptation with Representation Learning for Semantic Segmentation across Time
- Assia Benbihi, Matthieu Geist, Cédric Pradalier
- International Conference on Image Neural Processing 2019
- Learning Sensor Placement from Demonstration for UAV networks
- Assia Benbihi, Matthieu Geist, Cédric Pradalier
- IEEE Symposium on Computers and Communications 2019
- Automated segmentation and classification of land use from overhead imagery
- Antoine Richard, Assia Benbihi, Matthieu Geist, Cédric Pradalier, Vincent Perez, P Durand, R Van Couwenberghe
- International Conference on Precision Agriculture, 2018

Software

- WASABI (publication code)
- ELF (Publication code)
- Docker / Singularity makers Various Docker and Singularity definition file for 3D mapping and localization.