您最多选择25个主题 主题必须以中文或者字母或数字开头,可以包含连字符 (-),并且长度不得超过35个字符

1.8 KiB

Labeling

Accurately labeling assets with a predefined taxonomy will inform training and testing of algorithms as to which objects in a dataset have importance. Example: assets labeled with “table” and “chair” will provide an algorithm with the information it needs to train on identifying these objects separately within a scene.

You can add a Labeling component to individual GameObjects within a scene although it is a good practice to create a prefab of a GameModel and apply the Labeling component to it.

Multiple labels can be assigned to the same Labeling. When ground truth which requires unique labels per object is being generated, the first label in the Labeling present anywhere in the LabelingConfiguration is used.

Labeling Configuration

Many labelers require require a Labeling Configuration asset. This file specifies a list of all labels to be captured in the dataset for a labeler along with extra information used by the various labelers.

Best practices

Generally algorithm testing and training requires a single label on an asset for proper identification such as “chair”, “table”, or “door". To maximize asset reuse, however, it is useful to give each object multiple labels in a hierarchy.

For example An asset representing a box of Rice Krispies cereal could be labeled as food\cereal\kellogs\ricekrispies

  • “food” - type
  • “cereal” - subtype
  • “kellogs” - main descriptor
  • “ricekrispies” - sub descriptor

If the goal of the algorithm is to identify all objects in a scene that is “food” that label is available and can be used. Conversely if the goal is to identify only Rice Krispies cereal within a scene that label is also available. Depending on the goal of the algorithm any mix of labels in the hierarchy can be used.