Elahe Arani

Advanced Engineering (Czechia), Eindhoven University of Technology

Papers

Total Citations

H-Index

About

Elahe Arani is a leading researcher at the intersection of computer vision, adversarial machine learning, and continual learning, with a focus on robust 3D scene understanding. Her work primarily addresses the vulnerabilities and limitations of deep neural networks in real-world applications, particularly in monocular depth estimation and pose estimation. Arani’s major contributions include pioneering adversarial attacks on monocular pose estimation (8 citations), revealing critical security flaws in autonomous systems, and advancing continual learning for unsupervised monocular depth from videos (7 citations), enabling models to adapt to new environments without forgetting prior knowledge. She also developed image masking techniques for robust self-supervised monocular depth estimation (5 citations), improving accuracy under challenging conditions, and explored transformers in unsupervised structure-from-motion (2 citations), pushing the boundaries of 3D reconstruction. With over 20 total citations across her most-cited papers, Arani’s research is highly relevant for students and researchers in robotics, autonomous driving, and reliable AI deployment. Her notable achievements include bridging the gap between adversarial robustness and continual learning, offering practical solutions for safe and adaptive computer vision systems.