Falcor3D, Isaac3D
收藏资源简介:
Falcor3D数据集包含233,280张基于3D客厅场景的1024x1024分辨率图像,涉及7个变量因素。Isaac3D数据集包含737,280张基于3D厨房场景的512x512分辨率图像,涉及9个变量因素。两者均用于解耦学习的高级挑战。
The Falcor3D dataset comprises 233,280 images of 1024x1024 resolution based on 3D living room scenes, involving 7 variable factors. The Isaac3D dataset includes 737,280 images of 512x512 resolution based on 3D kitchen scenes, involving 9 variable factors. Both datasets are utilized for advanced challenges in decoupled learning.
数据集概述
Falcor3D
-
图像数量: 233,280
-
分辨率: 1024x1024
-
场景: 3D living room
-
变量因素: 7
- lighting_intensity (5 possible values)
- lighting_x-dir (6 possible values)
- lighting_y-dir (6 possible values)
- lighting_z-dir (6 possible values)
- camera_x-pos (6 possible values)
- camera_y-pos (6 possible values)
- camera_z-pos (6 possible values)
-
文件命名规则:
padded_index.png,其中index计算公式为:index = lighting_intensity * 46656 + lighting_x-dir * 7776 + lighting_y-dir * 1296 + lighting_z-dir * 216 + camera_x-pos * 36 + camera_y-pos * 6 + camera_z-pos
-
数据集大小: 98 GB
-
下采样版本: 128x128 (3.7 GB)
Isaac3D
-
图像数量: 737,280
-
分辨率: 512x512
-
场景: 3D kitchen
-
变量因素: 9
- object_shape (3 possible values)
- object_scale (4 possible values)
- camera_height (4 possible values)
- robot_x-movement (8 possible values)
- robot_y-movement (5 possible values)
- lighting_intensity (4 possible values)
- lighting_y-dir (6 possible values)
- object_color (4 possible values)
- wall_color (4 possible values)
-
文件命名规则:
padded_index.png,其中index计算公式为:index = object_shape * 245760 + object_scale * 30720 + camera_height * 6144 + robot_x-movement * 1536 + robot_y-movement * 384 + lighting_intensity * 96 + lighting_y-dir * 16 + object_color * 4 + wall_color
-
数据集大小: 190 GB
-
下采样版本: 128x128 (13 GB)
下载链接
许可证
- 许可证: Creative Commons Attribution 4.0 International License by NVIDIA Corporation




