The June sale is here! Save up to 70% on select products now through June 20.

3D Bounding Box for Deep Learning Object Detection

Plasma Labs - Code Plugins - Sep 13, 2023

Generate your 3D object detection dataset to train your own deep learning model.

  • Supported Platforms
  • Supported Engine Versions
    5.2 - 5.4
  • Download Type
    Engine Plugin
    This product contains a code plugin, complete with pre-built binaries and all its source code that integrates with Unreal Engine, which can be installed to an engine version of your choice then enabled on a per-project basis.

This plugin allows users to capture screenshots while extracting 3D bounding boxes around tracked objects, streamlining the annotation process.

Technical Details

Features:

  • 3D Bounding Boxes: Capture screenshots with accompanying label files that not only contain the coordinates of the 3D bounding boxes of tracked objects but also provide critical additional information:
  • Occlusion Detection: Automatically checks for object occlusion and annotates it in the label file if detected.
  • Truncation Check: Evaluates if objects are truncated and includes this information in the label, ensuring data accuracy.
  • Camera Visibility: Verifies if objects are in front of the camera; if not, they are excluded from the label file.
  • Camera calibration data: For each image generated, a file including the camera intrinsics matrix is saved.


Code Modules:

  • BoundingBox3D, Runtime

Important to know:

  • The plugin doesn't draw bounding boxes on your screen captures. It writes a .txt file containing the coordinates of the bounding boxes.


Number of Blueprints: 2

Number of C++ Classes: 5

Network Replicated: No

Supported Development Platforms:

  • Windows: Yes
  • Mac: Yes
  • Linux: No (not tested)

Supported Target Build Platforms:

  • Windows
  • Mac

Documentation: Plugin Documentation

Support Discord: Join Our Discord Community