Center of geometry#

Overview#

The ground truth exposed in this modality is the geometric center of each human or other important object in the scene.

This modality is not available in the Faces platform.

Relevant file	Location
center_of_geometry.json	camera folder

center_of_geometry.json#

This file is not provided in the Faces platform.

This file contains the location of the center of geometry of objects in the scene for which that information is available. It has the following format:

{
    "version": "1.0.0",
    "body_001": {
        "semantic_name": "human",
        "center_of_mass": {
             "global_3d": {
                 "x": 0.8170781433582306,
                 "y": 0.9888407624910058,
                 "z": 0.7662263974445049
             },
             "pixel_2d": {
                 "x": 125,
                 "y": 518,
                 "depth": 0.3274444043636322
             }
        }
    }
}

version: String. Version tracking for this file. Whenever you access this file in a datapoint, make sure to check that the version matches what you expect it to be; otherwise its format and fields may not be recognized.
body_001, body_002, etc.: Objects. One object for each item in the scene for which the center of geometry is available.
- semantic_name: String. An identifier of the type of object this is.
- center_of_mass: Object. A keypoint object.
  - global_3d: Object. Contains a set of three Floats giving the location of the keypoint in global coordinates: “x”, “y”, and “z”. See About our coordinate systems for details.
  - pixel_2d: Object. Contains a set of two Ints giving the x and y coordinates of the keypoint in the rendered image itself. See About our coordinate systems for details. In addition, this keypoint contains a unique depth field that gives the perpendicular distance, in meters, from the camera plane to the object’s center of geometry (in other words, the 3D Y coordinate of the center of geometry in the camera space).

Using this ground truth, you can train your model to recognize typical distances between the camera and objects in the scene.