Open mslavescu opened 5 years ago
The videos from this series have very good quality, 4K, and they cover a lot of different cities. Here is a good example:
London Drive 4K - Modern Day Westminster - UK https://youtu.be/hjzHjsvoi64
Let's run this video through the algorithms described here: https://medium.com/@mslavescu/try-live-ssd-object-detection-mask-r-cnn-object-detection-and-instance-segmentation-sfmlearner-df62bdc97d52 And analyze their output.
A large labeled crowdsourced dataset:
BDD100K: A Large-scale Diverse Driving Video Database
https://bair.berkeley.edu/blog/2018/05/30/bdd/ From the article: As suggested in the name, our dataset consists of 100,000 videos. Each video is about 40 seconds long, 720p, and 30 fps. The videos also come with GPS/IMU information recorded by cell-phones to show rough driving trajectories. Our videos were collected from diverse locations in the United States, as shown in the figure above
https://bair.berkeley.edu/blog/2018/06/18/bdd-update/ From the article: Importantly, different datasets focus on different aspects of the autonomous driving challenge. Our dataset is crowd-sourced, and covers a very large area and diverse visual phenomena (indeed significantly more diverse than previous efforts, in our view), but it is very clearly limited to monocular RGB image data and associated mobile device metadata. Other dataset collection efforts are complementary in our view. Baidu’s, KITTI, and CityScapes each contain important additional sensing modalities and are collected with fully calibrated apparatus including actuation channels. (The dataset from Mapillary is also notable, and similar to ours in being diverse, crowd-sourced, and densely annotated, but differs in that we include video and dynamic metadata relevant to driving control.) We look forward to projects at Berkeley and elsewhere that leverage both BDD100K and these other datasets as the research community brings the potential of autonomous driving to reality.
https://medium.com/@karol_majek/bdd100k-dataset-25e83e09ebf8
nuScenes open-sources self-driving dataset with 1.4M images https://www.therobotreport.com/nuscenes-self-driving-dataset/ https://www.nuscenes.org/ https://www.nuscenes.org/explore/scene-0011/0
nuTonomy released a self-driving dataset called nuScenes that it claims is the “largest open-source, multi-sensor self-driving dataset available to public.” According to nuTonomy, other self-driving datasets such as Cityscapes, Mapillary Vistas, Apolloscapes, and Berkeley Deep Drive focused only on camera-based object detection.
In March 2019, we released the full nuScenes dataset with all 1,000 scenes. The full dataset includes approximately 1.4M camera images, 390k LIDAR sweeps, 1.4M RADAR sweeps and 1.4M object bounding boxes in 40k keyframes. Additional features (map layers, raw sensor data, etc.) will follow soon. We are also organizing the nuScenes 3D detection challenge as part of the Workshop on Autonomous Driving at CVPR 2019.
https://www.nuscenes.org/overview The nuScenes dataset is inspired by the pioneering KITTI dataset. nuScenes is the first large-scale dataset to provide data from the entire sensor suite of an autonomous vehicle (6 cameras, 1 LIDAR, 5 RADAR, GPS, IMU). Compared to KITTI, nuScenes includes 7x more object annotations.
Collect YouTube, Vimeo etc videos links, with all kinds of driving, that are good for testing SDC algorithms especially in the area of LKAS, VisionBased ACC, FCW, Pedestrian avoidance, etc. You can try a few algorithms here, for free in the cloud with GPU acceleration, more will come soon:
https://medium.com/@mslavescu/try-live-ssd-object-detection-mask-r-cnn-object-detection-and-instance-segmentation-sfmlearner-df62bdc97d52
Here is my list of public videos as a starting point:
https://www.youtube.com/playlist?list=PLUop7b1Q1uZkv5__d2yPZG1cAXcelata8
Add your contributions in the comments here on this issue, or as comments on this video: https://youtu.be/Z3bxoi0ZJ_g