In fleet telematics, extracting scenarios from dashcam data is key for improving safety and efficiency. Paridhi’s talk focuses on developing a Perception-scenario extraction pipeline using a single-camera system for object detection and depth estimation. She’ll cover challenges like occlusion and scale regression, and share solutions for accurate modeling. Paridhi also highlights the versatility of Ridecell’s techniques across domains, enhancing applications in surveillance, autonomous vehicles, and robotics.