End-to-end pipeline for 3D scene reconstruction and visualization of autonomous driving footage using object detection, depth estimation, and semantic rendering in Blender. Integrates DETiC, Mask RCNN, ZoeDepth, and other models to generate realistic traffic scenes from monocular video.