Getting Started with Lidar Labeler App, Part 4: Labeling Automation Using Pretrained Deep Learning Models

This is the last video of a four-part series. It will describe a deep learning automation algorithm designed for computer vision tasks in 3D point clouds.

Users can utilize pretrained deep learning models or import their own custom-trained models for object detection. The algorithm supports both cuboid and semantic label outputs, providing flexibility and accuracy in labeling.

The automation algorithm maps the network output to the defined label definitions, streamlining the labeling process and producing precise results.