KITTI was co-founded by Karlsruhe Institute of Technology in Germany and the Toyota Institute of Technology in the United States. It is currently the world’s largest computer vision algorithm evaluation data set under auto-driving scenarios. It is used to evaluate the performance of computer vision technologies such as target detection (vehicle, non-motor vehicle, pedestrian, etc.) detection, target tracking, and road segmentation in an automotive environment. KITTI contains real image data collected from urban, rural, and highway scenes, with as many as 15 vehicles and 30 pedestrians in each image, as well as various degrees of shelter.
The Cityscapes dataset is driven by Mercedes-Benz and provides an image segmentation dataset in an unmanned environment. Used to evaluate the performance of visual algorithms in semantic understanding of urban scenes. Cityscapes contains 50 city scenes with different scenes, different backgrounds, and different seasons. It provides 5,000 fine-labeled images, 20,000 rough-labeled images, and 30 types of marked objects. The algorithm performance was evaluated using the PASCALVOC standard intersection-over-union (IoU) score.
In the latest KITTI and Cityscapes rankings, the performance of domestic startups is very eye-catching. In the KITTI dataset, the target detected three single items, the target tracked two single items, and the road splits four single items, etc., a total of nine single items. From the domestic Tucson all occupied the first place in the list, and Baidu participated in the evaluation. Samsung Institute, Invista, UCSD, NEC Laboratories, Stanford, University of Toronto and other famous manufacturers and research institutes. At the same time, on the Cityscapes data set, Tucson ranked first, and another domestic artificial intelligence startup was Shangyu Technology, followed by the Samsung Institute.