C.LABEL

Annotate smartly

Learn more

 

 

Designed to meet the requirements of even the most complex annotation projects in autonomous driving.

 

High quality training and test data are essential for the development of algorithms and functions for various automotive and industrial applications.

C.LABEL is an easy-to-useflexible and extensible software for the efficient annotation, quality control and correction of automotive data. These data can then be used for the development of autonomous driving algorithms and functions, as well as the testing and validation thereof.

With C.LABEL, we provide a semi-automated solution for the annotation of camera, LIDAR and RADAR data, as well as the possibility to combine different sensor data to improve accuracy.

Annotation methods:

  • Bounding box annotation for 2D images and 3D point cloud using rectangles and cuboids
  • Semantic segmentation of 2D images using polygons
  • Semantic segmentation in 3D point clouds using point clusters
  • Annotation of maximum object dimensions in 2D images using extreme points
  • Annotation of object position using key-points and 2D sub-bounding boxes (e.g. pedestrian pose)
  • Lane annotation in 2D images using splines

 

 User interface  Annotation methods

 

Additional key features:

  • User-friendly, swift and sleek UI
  • Flexible definition of object categories and properties
  • Support of various input and output formats
  • Semi-automated features for efficient annotation
  • Seamless integration with CMORE Automotive software solutions (e.g. C.FRAME, C.LYTIX, C.DATA)

Smart features

Semi-automated features for efficient annotation: 

  • Interpolation and extrapolation of bounding boxes, polygons, curves, key-points and extreme points
  • One-shot creation of 3D bounding boxes and semantic segmentation of point clusters
  • Automatic polygon generation from 2D bounding box
  • One-click pre-annotation of objects using 2D and 3D bounding boxes in camera images and point clouds

User-friendly and flexible UI

The user interface of C.LABEL is designed to minimize the effort of the user by providing special features and enabling a flexible configuration depending on individual needs.

Highlights:

  • Multi-View and tabbing functionality
  • Support of multiple screens
  • Flexible configuration of object types
  • Definition of global group and class properties
  • Definition of mandatory and optional properties
  • Various element types (checkbox, drop-down list, number, etc.) for defining properties
  • Creation and visualization of statistical reports

Multi-view and tabbing 

The multi-view and tabbing functionality allows the user to visualize several data streams simultaneously as well as to drag them freely into the tool frame or view on additional monitors.

Annotation methods in C.LABEL

Bounding Box Annotation

Semantic segmentation Annotation

Point Cloud Annotation

Benefits to your business

Your competitive advantage is that C.LABEL can be customized by our experienced engineers according to your requests.

We can help you with the clarification of label specifications and project requirements, so that your ground truth data suits your needs.

We provide you with additional cost-effective solutions for large labeling projects encompassing project management, data management, workforce and much more.

Check out our page Turn-key Labeling Solutions for Automotive Data and learn more about how we can support you.

Let‘s Get In Touch!

For customized requests and detailed information
we stay always at your disposal.