Skip to content

Complete Feature Set

At, we believe quality of the machine learning model is dependent on the quality of labelling performed on the training data. We empower data-scientists to control the quality of data labelling by providing highly configurable training data management system:

Segmentation and Classification's application supports following types of annotation work:

  1. Segmentation only
  2. Classification only
  3. Segmentation with Classification.

Pixel Accurate Tools

We use advanced image processing to provide pixel accurate tools like

  1. Superpixel-segmentation with brush and eraser,
  2. Polygon with advanced editing, bounding box,
  3. Freehand-draw with sculpter, and
  4. Region of Interest (ROI) growth tool for DICOM data.

AI Assisted Annotation

  1. AI Assisted Annotation using Data-scientist's ML model.
  2. AI Assisted Annotation using NVIDIA Clara for radiology.

Data Security and Privacy Controls

On-premises hosting of datasets and annotation tools using Docker and VPN.

Collaboration: Annotator's Performance Management

  1. Measure, record and analyze annotator's performance on every individual task, asset and label.
  2. Compare performance of multiple annotators on same task.
  3. Distribute labeling work among multiple labelers and observe consensus among their work.
  4. Seed annotation tasks with golden data set. Report performance of annotator on golden data set.

Labeling Instruction Builder

Labeling instruction builder empowers data scientist to define user experience of annotators in fine-detail:

  1. Labeling classes and ontology can be defined in fine-detail.
  2. WYSIWYG instruction builder enables data-scientist to view how labeling interface appears to the annotator.

We believe the quality of work performed by annotators depends on the quality of user experience in annotation tools. We build intuitive user experience for annotators.

WorkForce: Humans-in-the-Loop

We partner with highest quality workforce from around the world.