Playing for Benchmarks

We present a benchmark suite for visual perception. The benchmark is based on more than 250K high-resolution video frames, all annotated with ground-truth data for both low-level and high-level vision tasks, including optical flow, semantic instance segmentation, object detection and tracking, object-level 3D scene layout, and visual odometry. Ground-truth data for all tasks is available for every frame.

Released in: Playing for Benchmarks

Source: arXiv - Playing for Benchmarks

Contributor:

by

Homepage

Summary

We present a benchmark suite for visual perception. The benchmark is based on more than 250K high-resolution video frames, all annotated with ground-truth data for both low-level and high-level vision tasks, including optical flow, semantic instance segmentation, object detection and tracking, object-level 3D scene layout, and visual odometry. Ground-truth data for all tasks is available for every frame.

Uknown

Images in dataset

2017

Year Released

Key Links & Stats

@InProceedings{Richter_2017, title = {Playing for Benchmarks}, author = {Stephan R. Richter and Zeeshan Hayder and Vladlen Koltun}, booktitle = {{IEEE} International Conference on Computer Vision, {ICCV} 2017, Venice, Italy, October 22-29, 2017}, pages = {2232--2241}, year = {2017}, url = {https://doi.org/10.1109/ICCV.2017.243}, doi = {10.1109/ICCV.2017.243}, }

scenebox

Modalities

  1. Video

Verticals

  1. A/V

ML Task

  1. Image Classification
  2. Object Detection
  3. Autonomous Vehicles

Related organizations