Vision-based vehicle distance estimation provides a scalable solution to safety challenges in autonomous driving, particularly in scenarios where RADAR or LiDAR sensors may be ineffective. This project develops a forward-vision pipeline that leverages object detection and monocular depth estimation on camera inputs, validated using CARLA simulations and real-world Argonne datasets. A hybrid ensemble combining geometric and deep learning approaches demonstrates strong performance and real-world adaptability.
Watch the team present this project at 32:35 in the session recording here.
Keywords: Computer Vision, Visual Transformers, Deep Learning, Autonomous Vehicles, Monocular Depth Estimation, Digital Image Processing, Real-time Inference, PII Filtering, Image Segmentation, Convolutional Neural Networks, Zero-Shot Inference
Faculty Advisor
PhD in Theoretical Physics, MS in Computer Science. Currently – Applied Scientist at Amazon. Previous jobs: Computational Scientist at the Argonne National Laboratory, Scientist at LIGO project of California Institute of Technology. I specialize in scientific computing, HPC, machine learning, data analysis.