KITTI-360-APS

displayName: KITTI-360-APS
license:
- KITTI-360-APS Custom
paperUrl: https://openaccess.thecvf.com/content/CVPR2022/papers/Mohan_Amodal_Panoptic_Segmentation_CVPR_2022_paper.pdf
publishDate: "2022"
publishUrl: http://amodal-panoptic.cs.uni-freiburg.de/
publisher:
- University of Freiburg
tags:
- Pedestrians
- Vehicles
- Roads
- Traffic Sign

---
# 数据集介绍
## 简介
我们扩展了KITTI 360数据集，该数据集具有带有amodal panoptic注释的语义和实例标签，并将其命名为KITTI-360-APS数据集。它由九个城市街道场景序列组成，并为分辨率1408x376像素的61,168图像提供注释。我们的数据集包括10个类。如果类具有无定形区域或在任何时间点都无法移动，则我们将该类定义为东西。道路，人行道，建筑物，墙壁，围栏，杆子，交通标志，植被，地形和天空是这些东西。此外，数据集由7个事物类别组成，即汽车，行人，骑自行车的人，两轮车，货车，卡车和其他车辆。请注意，我们将自行车和摩托车类合并为一个称为两轮车的单一类。

## Download dataset
:modelscope-code[]{type="git"}

displayName: KITTI-360-APS
license:
- KITTI-360-APS Custom
paperUrl: https://openaccess.thecvf.com/content/CVPR2022/papers/Mohan_Amodal_Panoptic_Segmentation_CVPR_2022_paper.pdf
publishDate: "2022"
publishUrl: http://amodal-panoptic.cs.uni-freiburg.de/
publisher:
- University of Freiburg
tags:
- Pedestrians
- Vehicles
- Roads
- Traffic Sign

---
# Dataset Introduction
## Overview
We extend the KITTI 360 dataset, which is equipped with semantic and instance labels paired with amodal panoptic annotations, and name this extended dataset KITTI-360-APS. It comprises nine urban street scene sequences, and provides annotations for 61,168 images with a resolution of 1408×376 pixels. Our dataset encompasses 10 categories. We define a class as "stuff" if it has an amorphous region or is immovable at any time point. Roads, sidewalks, buildings, walls, fences, poles, traffic signs, vegetation, terrain, and sky fall into this stuff category. Additionally, the dataset consists of 7 "thing" categories, namely cars, pedestrians, cyclists, two-wheelers, vans, trucks, and other vehicles. Note that we merge the bicycle and motorcycle classes into a single class called two-wheelers.

## Download Dataset
:modelscope-code[]{type="git"}