Apply 3D Detection to Pointcloud Project

Run 3D Detection and tracking algorithm on pointclouds or pointcloud episodes project

Apply 3D Detection to Pointcloud Project

OverviewUsageHow To RunResultRelated AppsRelated ProjectsScreenshot

GitHub release (latest SemVer) views runs

Overview

This app provides convenient inference interface for 3d neural networks on pointclouds. Just connect to your deployed model, select model classes and model predictions will be saved to a new project.

Application key points:

  • Works with both separate pointclouds and pointcloud episodes
  • Support 3d tracking

Usage


How To Run

Step 0. Please make sure that you deployed 3d detection model using corresponding serving app. Learn more in ecosystem …

Step 1. Add Apply 3D Detection to Pointcloud Project application to your Team

Step 2. Launch from Pointcloud Project's context menu

Step 3. Press the Run button in the modal window

Result

Related Apps

  1. Train MMDetection3D - start training on your custom data. Just run app from the context menu of your project, choose classes of interest, train/val splits, configure training parameters and augmentations, and monitor training metrics in realtime. All training artifacts including model weights will be saved to Team Files and can be easily downloaded.
  1. Serve MMDetection3D - serve model as Rest API service. You can run pretrained model, use custom model weights trained in Supervisely. Thus other apps from Ecosystem can get predictions from the deployed model. Also developers can send inference requiests in a few lines of python code.
  1. Import KITTI 3D - app allows to get sample from KITTI 3D dataset or upload your downloaded KITTI data to Supervisely in point clouds project format.
  1. Import KITTI-360 - app allows to upload your downloaded KITTI-360 data to Supervisely in point cloud episodes project format.

Related Projects

  1. Demo LYFT 3D dataset annotated - demo sample from Lyft dataset with labels.
  1. Demo LYFT 3D dataset - demo sample from Lyft dataset without labels.
  1. Demo KITTI pointcloud episodes annotated - demo sample from KITTI 3D dataset with labels.
  1. Demo KITTI pointcloud episodes - demo sample from KITTI 3D dataset without labels.

Screenshot