Data Labeling Standard
Labeling quality is one of the most significant factors affecting model performance. In actual projects, low labeling quality accounts for the reasons for more than 90% of poor model performance cases. Therefore, if the model is not performing well, solving labeling quality issues should be prioritized. This topic describes the labeling methods for different types of objects and the requirements for labeling quality.
Determine the Labeling Method
-
Label the upper surface’ contour
-
Objects: Regular objects that are laid flat, such as cartons, medicine boxes, rectangular objects, etc.
-
Labeling tools: Rectangle Tool and Polygon Tool.
-
Labeling method: The pick points are calculated on the upper surface contour, and the user only needs to make rectangular selections on the images.
-
-
Label the entire objects’ contours
-
Objects: Sacks, various types of objects, etc.
-
Labeling tool: Polygon Tool.
-
Labeling method: Labeling the object contours is the general method, which is suitable for most objects.
-
-
Special cases when the recognition result needs to conform to how the grippers work
-
Bottles:
-
Scenario: It is necessary to ensure that the suction cup and the tip of the bottle to pick completely fit (high precision is required).
-
Labeling method: Only the bottle tip contours need to be labeled.
-
-
Rotors:
-
Scenario: The task of rotor picking involves recognizing rotor orientations.
-
Labeling method :Only the middle parts whose orientations are clear can be labeled, and the thin rods at both ends cannot be labeled.
-
-
Metal pieces:
-
Scenario: It is necessary to ensure that the suction parts are in the middle parts of the metal pieces.
-
Labeling method: Only the middle parts of the metal pieces are labeled, and the ends do not need to be labeled.
-
-
Check Labeling Quality
The labeling quality should be ensured in terms of completeness, correctness, consistency, and accuracy:
-
Completeness
All objects that meet the requirement should be labeled. It is prohibited to miss an object that needs to be labeled.
-
Correctness
Make sure that the objects correspond to the class labels correctly. It is prohibited to assign a wrong label to the object.
-
Consistency
All data should be labeled with a consistent standard. For example, if a labeling rule stipulates that only objects that are over 85% exposed in the images be labeled, then all objects that meet the rule should be labeled. Please avoid situations where one object is labeled but another similar object is not.
-
Precision
The labeled contour should closely fit the edges of the object. Please avoid missing any object parts, or including excess regions outside the object contours.
