AI-Powered Data Annotation
Save time & resources with smart tools for polygon selection, frame-to-frame predictive labeling, & automated object recognition.
Get Started FREE
No Credit Card Required
SmartPoly Object Selection
Save time labeling polygons for instance segmentation by drawing a simple rectangle that resolves to an object.
TrackForward Labeling
Track a labeled objects from frame-to-frame to automatically apply polygons or bounding boxes.
AutoLabel Common Classes
Select common objects to recognize and Sense will use pre-trained models to automatically apply labels.
Label Your Data Faster, Smarter, Better
A full set of features to equip all your computer vision labeling projects. Sense Data Annotation is designed to maximize productivity and scale.
1. Ingest Data
2. Define Labels
Configure labels for your dataset. Select from: Rectangle, Polygon, Point, Feature Points, Text, Classification
3. Label Data
Use AI-Powered tools for fast & easy smart polygon selection, frame-to-frame object tracking, & common object auto-labeling
4. Export Datasets
Export labeled data to the format you need including COCO, Pascal VOC, YOLO, Create ML, or Sense’s own flexible JSON format
Sixgill is Trusted by Industry Leaders & Innovators:
Flexible Label Types
Sense supports all major definitions used in computer vision labeling. Choose from Rectangles, Polygons, Point, Feature Points, Text, Class, or Multi-Class.
Rectangles
Draw boxes to detect specific objects within images.
Polygons
Segment objects for detection within image frames.
Feature Points
Use for highly detailed detection such as emotion & body pose.
Text
Leverage for free text production for image captions ideal for accessibility.
Classification
Apply for image classification & attribute identification.
Sub-Labels
Add sub-labels into the label definitions for rectangles, polygons, and point labels.
Easy Exports
Sense Data Annotation supports the most common label formats for computer vision. Choose to export as COCO, Pascal VOC, YOLO, Create ML or Sense’s own JSON format.
COCO
A large-scale object detection, segmentation, and captioning dataset
Pascal VOC
Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes
YOLO
“You Only Look Once”. A popular real-time object detection algorithm
Create ML
Apple’s machine learning model creation and training framework
Sense JSON
Sense’s format is a schema defined JSON document containing all Sense label elements. Sense also includes metadata about the image and source video if applicable.
Team Collaboration
Sense Data Annotation makes labeling as a team easy. Add new members, sync data sources across teams, manage roles, provide labeling instructions, add comments, review labeling progress and time spent per label.
Cloud Sync
Sync data sources across teams
Instructions
Add labeling and project details
Project Review
Approve or reject labels & track progress
Data Source Connections
Connect both local storage and remote data sources to projects.
Supported image formats: JPG, PNG, GIF, BMP, TIFF, WEBP
Supported video formats from S3 & GCS : AVI, FLV, MKV, MOV, MP4, OGG, WEBM
Amazon S3
Connect an Amazon S3 bucket
Google Cloud Storage
Connect a GCP bucket
Local Assets
Upload files for labeling
Local CSV
Use a CSV file to upload assets
Get Started for FREE
No matter the size of your labeling projects, if you’re an expert, or just getting started with machine learning, Sense provides the speed, automation, and ease you need.