nyu_depth_v2

  • توضیحات :

مجموعه داده‌های NYU-Depth V2 از توالی‌های ویدئویی از انواع صحنه‌های داخلی تشکیل شده است که توسط دوربین‌های RGB و Depth مایکروسافت کینکت ضبط شده‌اند.

شکاف مثال ها
'train' 47584
'validation' 654
  • ساختار ویژگی :
FeaturesDict({
    'depth': Tensor(shape=(480, 640), dtype=float16),
    'image': Image(shape=(480, 640, 3), dtype=uint8),
})
  • مستندات ویژگی :
ویژگی کلاس شکل نوع D شرح
FeaturesDict
عمق تانسور (480، 640) float16
تصویر تصویر (480، 640، 3) uint8

تجسم

  • نقل قول :
@inproceedings{Silberman:ECCV12,
  author    = {Nathan Silberman, Derek Hoiem, Pushmeet Kohli and Rob Fergus},
  title     = {Indoor Segmentation and Support Inference from RGBD Images},
  booktitle = {ECCV},
  year      = {2012}
}
@inproceedings{icra_2019_fastdepth,
  author    = {Wofk, Diana and Ma, Fangchang and Yang, Tien-Ju and Karaman, Sertac and Sze, Vivienne},
  title     = {FastDepth: Fast Monocular Depth Estimation on Embedded Systems},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year      = {2019}
}