3D Point Cloud labeling annotation tool2024-03-21T13:39:32+01:00

Get your 3D data labeled faster

  • Automatically track and label objects in motion
  • Easily review and correct existing labels
  • Increase speed and accuracy with ML-assisted tooling

Trusted by computer vision engineers across robotics and AV companies

Label 3D point clouds faster

Create smart workflows and annotate your lidar and RGBD data easily. Segments.ai lets you visualize a point cloud and 2D images at the same time, so you can accurately label data in record time.

We support:

  • Cuboids
  • Keypoints
  • Polygons
  • Polylines
  • Segmentation

Labeling on Segments.ai is so quick and simple. Plus, the documentation is very good. It does everything we need, so we don’t have to worry about it.

Libor Novak

Software Engineer


Identify and classify objects accurately

Combine point clouds with visual camera data to make sure you’ve classified every object correctly. You’ll be able to instantly spot if every object was annotated correctly, and easily differentiate between objects.


Label dynamic objects quickly

Draw a box on the first frame, then let Batch Mode automatically track the object across frames. Instead of drawing hundreds of boxes, your team can focus on data quality.

When you need to correct a label, just update the cuboid in a single row. Segments.ai’s Keyframe Interpolation feature will automatically update that same object across every frame.


Label static objects more accurately

Get tight-fitting cuboid labels from the start. Turn on Merged Point Cloud Mode, and Segments.ai will merge all frames across the sequence into one big point cloud.

By aggregating the points across frames, the outlines of the static objects become clearly visible, allowing for accurate annotation with a tight-fitting cuboid in one go.


Track and label objects in motion

Speed up the labeling process. Keyframe Interpolation saves you time by automatically filling in an object’s position between two frames, while Smart Cuboid Propagation lets you label a moving object once, then automatically track and label it across frames.

Instead of filling in every frame manually, you’ll just have to verify and adjust the generated labels.


Identify different objects easily

Display point clouds with color, intensity values, or color by point height so you can easily differentiate between various objects and areas in your data.

Whether you’re working with lidar scans, radar, or any other point cloud data, you’ll be able to spot objects and label them faster and with more accuracy.


Upload point clouds of unlimited size

Speed up the process of loading whole scans, no matter how big the file size. The Segments.ai 3D Point Cloud Interface splits point clouds into 3D tiles, then only loads the visible tiles and nearby ones while you’re using the interface.

You’ll save time loading scans, make it easier on your GPU, and keep the labeling experience smooth.

Aggregate static data points across frames with merged point cloud mode


Rotate a cuboid on all 3 axes

Make it easy to annotate objects on a slope. Segments.ai lets you rotate the cuboid on all 3 axes for total labeling flexibility.

You’ll have better control over the orientation of the object, so you can make sure it’s labeled accurately.

And much more …

Segments.ai helps machine learning teams and data engineers move from POC to production, faster.
When you’re ready to scale your labeling process and make it more accurate and efficient, we’re here to help.

Annotation types2023-10-24T21:32:00+02:00

Segments.ai supports several different annotation types:

  • Semantic segmentation
  • Instance segmentation
  • Panoptic segmentation
  • Cuboids
  • Polygon
  • Polyline
  • Keypoint
Customizable hotkeys2023-10-24T21:33:50+02:00

Make labeling even faster and more efficient with customizable hotkeys. You can use hotkey shortcuts to:

Rotate, pan, move, and scroll
Zoom in on a view
Define rotation points for an object
Switch to an orthographic top view
Create and delete cuboids.

For more information on hotkeys, check out the documentation.

Dataset search syntax2023-10-24T21:34:47+02:00

Segments.ai makes it easy to find samples, fast. You can search for samples by name, metadata attributes, or label content.

You can learn more about the data search syntax in this documentation.

Keep the dimension of a cuboid the same across the whole sequence2023-10-24T21:35:26+02:00

Force objects to maintain the same dimensions in every frame. When changing the dimensions of an object with a certain track ID in one frame, the dimensions of the object will be updated in all other frames too.

This can be useful when you want objects to keep their dimensions, regardless of occlusions or noise.

Label on aligned camera images2023-07-03T14:57:42+02:00

Segments.ai allows you to upload images along with point clouds.

Additionally, you can specify the camera calibration parameters, allowing you to view the 3D point cloud and objects on the image.

Label sequences with interpolation2023-10-24T21:36:01+02:00

Keyframe Interpolation is the process of creating intermediate values between a set of keyframes. Segments.ai will automatically fill in the values between the selected keyframes.

Keyframe Interpolation is used for object labels that change over time. For example, say you have a sequence with a moving car, meaning the bounding box around the car will move in each frame. With Keyframe Interpolation, you only need to specify the label at certain key times, and the intermediate labels will be created automatically.

Select different color modes2023-10-24T21:37:07+02:00

The Object Coloring toggle lets you change how different objects in the sample are displayed. You can choose between two different displays:

By instance

  • Each object is displayed in a different color
  • E.g. two cars each get a different color

By category

  • Each category has a color defined in the dataset settings. Every object of the same category is displayed in that category color
  • E.g. two cars are displayed in the same category color
Toggle between different 3D panning modes2023-07-03T14:57:24+02:00

By default, map panning is enabled, which means the camera pans along the ground plane. E.g. when you pan up by clicking and dragging, or press w, the camera moves forwards.

When screen panning is enabled, the camera moves along the screen (i.e. along the camera plane). E.g. when you pan up by clicking and dragging, or press w, the camera moves up.

You can change the panning mode in the 3D interface settings.

Segments.ai had all the features required for us to label our images and point clouds, such as overlaying 3D point clouds onto 2D images. This proved very useful at giving labelers context for what they were looking at in 3D, especially in noisy shots.

Sarina H.

Robotics Software Engineer
OCIUS Technology Limited

Picture of someone labeling multi-sensor data on a point cloud

Labeling Services

Managing labeling workforces, whether in-house or external, is easy with Segments.ai.

If you’re unsure about where to start, we’re here to help. We have an extensive network of partners who we wholeheartedly trust with labeling.

A few of the labeling agencies we work with

Simplify multi-sensor labeling

Say hello to faster labeling, better quality data, and more time for your engineering team.
You’ve got access to a 14-day free trial to test it for yourself.

Go to Top