Segments.ai
Data labeling platform for robotics and autonomous driving that annotates 3D point clouds and camera images together.
Segments.ai is a data labeling platform for machine learning teams in robotics and autonomous driving. Segments.ai lets annotators label 3D point clouds and 2D camera images in the same task, keeps object track IDs consistent across sensors and time, and speeds up work with model-assisted tools. Segments.ai connects to training pipelines through a Python SDK and API.
Best for Computer vision teams that train perception models on combined lidar and camera data and need consistent labels across both.
- Listed in
- Image Recognition & OCRDeveloper API
- Platforms
- Web, API
- Pricing model
- Contact sales
- Free trial
- Yes
- API
- Yes
- Added
Last verified: . Information taken from segments.ai.
What is Segments.ai?
Segments.ai is an annotation platform designed for sensor data from robots and vehicles. Its main idea is multi-sensor labeling: lidar point clouds and camera frames from the same scene are opened together, an object receives one track ID in every modality, and a cuboid drawn in 3D can be projected onto the matching images. For point clouds the platform supports segmentation, cuboids, keypoints, polygons and polylines, with a batch view that lines up every appearance of an object in a sequence and a merged view that stacks frames so static objects are easier to outline. Interpolation between keyframes and machine-learning tracking reduce repeated work. For images there are superpixel, box-prompted and single-click segmentation tools, and predictions from your own model can be uploaded as pre-labels for correction. A Python SDK, API, webhooks and exports to common ML frameworks tie the tool into existing pipelines, and labeling can be outsourced to partner workforces.
Key Features
Multi-Sensor Labeling
Shows 2D images and 3D point clouds side by side so one annotation task covers every sensor in the scene.
3D-to-2D Projection
Copies a cuboid drawn in the point cloud onto each camera image where the object appears.
Batch Mode
Lists all instances of a tracked object on one scrollable screen; a correction in one frame carries over to the others.
Merged Point Cloud
Combines the frames of a sequence into a single dense cloud so sparse stationary objects can be boxed accurately.
ML-Assisted Image Tools
Offers adjustable superpixels, an autosegment tool driven by a drawn box, and hover-and-click masks based on the Segment Anything Model.
SDK, API and Exports
Integrates through a Python SDK, API keys and webhooks, and exports datasets for PyTorch, TensorFlow and Hugging Face.
Use Cases
- Autonomous Driving Datasets
- Annotate lidar and camera recordings together for perception model training.
- Delivery and Off-Road Robots
- Label sequences captured by mobile robots operating outside structured roads.
- Correcting Model Predictions
- Upload pre-labels from a trained model and have annotators verify and fix them.
- Tracking Moving Objects
- Label an object once and propagate the cuboid through a sequence with interpolation or tracking.
- Outsourced Annotation
- Bring in a partner labeling workforce while keeping quality control in the same tool.
Things to know
- Quotes are based on labeling hours
- Each plan is sized by a yearly amount of labeling usage rather than by seats, and seats are unlimited.
- Point cloud size depends on the plan
- The Core plan limits point clouds to 500,000 points each, while Fusion removes the size limit.
- Sensor fusion starts at Fusion
- The interfaces that combine several sensors are listed under the Fusion plan and above.
- Academic licenses
- Students and researchers can apply for a free academic license.
FAQs
What is Segments.ai?
Segments.ai is a multi-sensor data labeling platform used to annotate point clouds and images for robotics and autonomous vehicle machine learning.
What is Segments.ai used for?
Segments.ai is used to create training data for computer vision models, with labels such as cuboids, segmentation masks, bounding boxes, polygons, polylines and keypoints.
Is there a free trial of Segments.ai?
Yes. Segments.ai offers a 14-day free trial.
How much does Segments.ai cost?
Segments.ai lists three plans, Core, Fusion and Enterprise, each with a custom quote based on yearly labeling hours. See the official pricing page for details.
Does Segments.ai have an API?
Yes. Segments.ai provides an API and a Python SDK, and the Fusion plan adds a webhooks system.
Is Segments.ai free for researchers?
Segments.ai says it offers free academic licenses to students and researchers, and is open to discussing options with startups.
Does Segments.ai offer labeling services?
Yes. Segments.ai connects customers with partner labeling agencies for teams that prefer to outsource annotation.
Who it's for
- Computer vision engineers
- Machine learning teams
- Robotics companies
- Autonomous vehicle teams
- Academic researchers
Comments
No comments yet. If you have used Segments.ai, tell other readers how it went.
Alternatives to Segments.ai
| Tool | Free plan | Starting price | API |
|---|---|---|---|
 CVATData annotation platform for labelling images, video, 3D point clouds and audio, offered as cloud, self-hosted and open source. |
Yes | — | Yes |
 Label StudioOpen source platform for labeling data and evaluating AI models with human review. |
Yes | — | Yes |
UltralyticsComputer vision platform from the makers of YOLO for annotating data, training models on cloud GPUs and deploying them. |
Yes | — | Yes |
 Digitap.aiAPI platform for banks and fintechs covering digital KYC onboarding, bank statement analysis and alternate-data credit assessment. |
— | — | Yes |
 DatatureNo-code computer vision platform for labelling image data, training models and deploying them to cloud or edge. |
Yes | — | Yes |
Listed in the same categories as Segments.ai. A starting price appears only where we checked it on the maker’s pricing page; a dash means not checked or not published. All Image Recognition & OCR tools