How many cars are visible in a given image? Is the sky clear or cloudy? SynScapes provides a wide range of metadata which helps characterize each image.

Additionally, SynScapes offers information that is difficult for humans to annotate, such as accurate per-instance visibility.
 

dataset Layout

SynScapes is organized into the following directories:

├── img
│   ├── class     [1-25000].png
│   ├── depth     [1-25000].exr
│   ├── instance  [1-25000].png
│   ├── rgb       [1-25000].png
│   └── rgb-2k    [1-25000].png
└── meta          [1-25000].json

Image resolution

SynScapes' native resolution is 1440x720, stored in the img/rgb folder. In order to best support training with architectures designed for Cityscapes, we also include an up-scaled version at 2048x1024 resolution in img/rgb-2k. Note that this up-scaling precedes the sensor simulation stage, ensuring pixel noise is present at the appropriate scale.
 

Annotations

  RGB

RGB

  Class  as single-channel PNG (visualized in color above)  The class annotations follows  the Cityscapes convention . 

Class as single-channel PNG (visualized in color above)

The class annotations follows the Cityscapes convention

  Instance  as PNG  The instance id can be found as  R + G * 256 + B * 256^2 .

Instance as PNG

The instance id can be found as R + G * 256 + B * 256^2.

  Depth  as floating point OpenEXR  Stores the planar depth (not distance) in meters.  

Depth as floating point OpenEXR

Stores the planar depth (not distance) in meters.
 

Camera metadata

The camera's position and field of view is as follows:

"camera": {
    "extrinsic": {
      "pitch": 0.038, 
      "roll": -0.0, 
      "x": 1.7, 
      "y": 0.1, 
      "yaw": -0.0195, 
      "z": 1.22
    }, 
    "intrinsic": {
      "fx": 1590.83437, 
      "fy": 1592.79032, 
      "resx": 1440, 
      "resy": 720, 
      "u0": 771.31406, 
      "v0": 360.79945
    }
  } 

Instance metadata

  2D Bounding Boxes

2D Bounding Boxes

  3D Bounding Boxes  in ego-vehicle coordinates

3D Bounding Boxes in ego-vehicle coordinates

  Occlusion  (fraction of object hidden behind other objects)

Occlusion (fraction of object hidden behind other objects)

  Truncation  (fraction of object outside field of view)

Truncation (fraction of object outside field of view)

Note: 'class' is also recorded in the JSON file, to facilitate instance-to-class mapping without having to refer to the PNG file.
 

scene metadata

altitude_variation 
The largest altitude difference in the scene in meters.

curb_height 
The height of the sidewalk curb in meters.

dist_*_{mean,stddev}
For each actor class, contains the mean and standard deviation of distance for all visible instances.

ego_speed
The speed in m/s traveled by the ego vehicle at the time of image capture.

fence_{presence,height}
Indicates whether fences are present in the image. Note that due to occlusion, it may be hidden behind another object. Height is measured in meters.

median_presence
Whether the road median is present.

num_*
For each actor class, contains the number of visible instances.

parking_{presence,angle}
Whether a parking lane is present, and whether cars park at 0 (parallel), 45 or 90 degrees.

rel_dist_to_isect
Relative distance to nearest intersection. 0.0 indicates ego vehicle is inside the intersection, 1.0 indicates it is one city block away from the next intersection.

road_material_type
Integer representing the material used for the road surface.

sidewalk_width
The width of the sidewalk in meters

sky_contrast
Contains the logarithm of the sky's contrast, measured as max/mean. Values around 2-3 indicate fully overcast sky, 5-6 indicate direct sunlight.

sun_height
The normalized angular height of the sun. 0.0 indicates sunset/sunrise, 1.0 indicates zenith.

wall_{presence,height}
Whether the wall class is present, with height in meters.