This task aims to extract depth information from a 2D image using internal camera characteristics, also known as the Intrinsic Matrix (IM). The dataset we are using is a selectively sampled subset from the KITTI dataset (provided in the repo).
The calculated distances will be compared with the distances provided in the dataset labels. We are using the YOLOv5s. object detection model to detect and locate cars in the scene.