KittiViewer-plus | KittiViewer+ is an upgrade of Second.KittiViewer

It's easier to use conda to get QT:.
kitti_detection_root parameter contains a path for detection KITTI dataset. You must have image_2, label_2, velodyne, calib folders inside "selected_split": "training" folder. folders parameter contains names for each modality (if you named them differently from KITTI). compare_results_detection object describes how to display results from different methods. All results must be stored inside the root folder and methods[i]["name"] subfolder in KITTI annotations format (.txt files as in label_2). color field describes RGB color of 3d and 2d boxes.
kitti_detection_root parameter contains a path for detection KITTI dataset. You must have image_2, label_2, velodyne, calib folders inside "selected_split": "training" folder. "kitti_detection_root": "/data/sets/kitti_second"
folders parameter contains names for each modality (if you named them differently from KITTI). "folders": { "detection": { "calib": "calib", "image": "image_2", "label": "label_2", "velodyne": "velodyne" }, }
compare_results_detection object describes how to display results from different methods. All results must be stored inside the root folder and methods[i]["name"] subfolder in KITTI annotations format (.txt files as in label_2). color field describes RGB color of 3d and 2d boxes. "compare_results_detection": { "root": "/data/sets/kitti_results_detection", "methods": [ { "name": "method1", "color": [1.0, 0.0, 1.0] }, { "name": "method2", "color": [1.0, 1.0, 0.0] }, { "name": "method3", "color": [0.0, 0.0, 1.0] } ] }
kitti_tracking_root parameter contains a path for detection KITTI dataset. You must have image_2, label_2, velodyne, calib folders inside "selected_split": "training" folder. image_2 and velodyne must have folders for each scene. folders parameter contains names for each modality (if you named them differently from KITTI). compare_results_tracking object describes how to display tracking results from different methods. All results must be stored inside the root folder and methods[i]["name"] subfolder in selected format. color field describes RGB color of 3d and 2d boxes.
kitti_tracking_root parameter contains a path for detection KITTI dataset. You must have image_2, label_2, velodyne, calib folders inside "selected_split": "training" folder. image_2 and velodyne must have folders for each scene. "kitti_tracking_root": "/data/sets/kitti_tracking"
folders parameter contains names for each modality (if you named them differently from KITTI). "folders": { "tracking": { "calib": "calib", "image": "image_2", "label": "label_2", "velodyne": "velodyne" }, }
compare_results_tracking object describes how to display tracking results from different methods. All results must be stored inside the root folder and methods[i]["name"] subfolder in selected format. color field describes RGB color of 3d and 2d boxes. "compare_results_detection": { "root": "/data/sets/kitti_results_detection", "format": "ab3dmot", "methods": [ { "name": "method1", "color": [1.0, 0.0, 1.0] }, { "name": "method2", "color": [1.0, 1.0, 0.0] }, { "name": "method3", "color": [0.0, 0.0, 1.0] } ] }
Supported tracking annotations formats: ab3dmot: comma separated values in order:
kitti_segmentation_root parameter contains a path for detection KITTI dataset. You must have .bin files "selected_split": "training" folder. Each file is created using dill library and must contain segmented_points (NxC, C - amount of classes (currently C=1 only),original_points (Nx3 - with points coordinates), ground_trurh_points (NxC - with ground truth segmentation).
kitti_segmentation_root parameter contains a path for detection KITTI dataset. You must have .bin files "selected_split": "training" folder. Each file is created using dill library and must contain segmented_points (NxC, C - amount of classes (currently C=1 only),original_points (Nx3 - with points coordinates), ground_trurh_points (NxC - with ground truth segmentation) "kitti_segmentation_root": "/data/sets/kitti_segmentation"