Gastroenterology
AI is reducing bias in cancerous polyp detection and inflammatory bowel disease.
- Track polyps in video
- Colonography and scope videos in one tool
- Support Mayo and UCEIS scoring rubrics
Accelerating AI in healthcare
Encord empowers leading medical institutions and companies to address some of the hardest challenges in computer vision for healthcare.
Use cases
Transforming healthcare for better patient outcomes
Computer vision is revolutionising progress in medical imaging - from diagnostics and pathology to robotic surgery.
Radiology
Annotate modalities such as CT, MRI, and X-ray.
- High-precision annotation of voxels
- Hanging protocols
- Axial, coronal, and sagittal view
Microscopy
Handle high-resolution microscopy images with ease.
- Image processing and filtering options
- Automated measuring capabilities
- Tools for fine-grained segmentations
Ultrasound
Annotate ultrasound data with high accuracy.
- Expert review workflows
- Object detection, classifications and segmentations
- Handle 2D and 3D
Dermatology
Label and classify a variety of conditions.
- Flexible label taxonomies
- Pixel perfect segmentation of lesions
- Custom review processes keep experts in the loop
Surgical devices
Accurately track devices in surgical videos.
- Native video capabilities for streamlined annotation
- Automation features reduce manual burdens
- 100% data-secure for privacy guarantees
Stanford Medicine
The Division of Nephrology reduced experiment duration by 80% while processing 3x more images.
Problem
Stanford was using three different pieces of software to identify, annotate, and count podocytes and glomeruli in microscopy images.
Solution
Stanford started using Encord’s annotation tools & SDK to automate segmentations, count, and calculate sizes of segments.
Results
With Encord, Stanford researchers reduced experiment duration from an average of 21 to 4 days while processing 3x the number of images.
80%
Reduction in experiment duration
3X
Number of images
1 platform
... and not 3
How it works
Emphasising data quality for healthcare
Expert review workflows, fully auditable, automated data quality checks. Fully compliant medical-specific tooling.
Specialised tools
- Annotate medical imaging data such as microscopy images, ultrasound videos, and volumetric DICOM images, with one platform.
Expert review
- Define custom expert review workflows to make the most of your domain experts.
Automated quality assessment
- Combine human specialists with automated intelligence to ensure the highest quality standard for your data and labels.
Automated labelling
- Use our object tracking & interpolation features to reduce costs. Use micro-models to accelerate your active learning workflows and get to production faster.
Compliant & secure
- Capture maximal efficiency gains of our automation features while retaining 100% control of your data.
Auditable
- Ensure better, faster, and cheaper approval processes. Closely monitor annotator throughput and quality.
Native medical imaging capabilities
We support all common medical image modalities including ultrasound, gastro videos, and offer a leading solution for volumetric DICOM images.
Granular quality control
From satisfying regulatory and compliance requirements to offering customisable annotation review processes. Our software is developed alongside, and designed for the use by medical experts.
King's College London
KCL used Encord to achieve a 6.4x average increase in labelling efficiency for GI videos.
Problem
Using clinicians to annotate pre-cancerous polyp videos had prohibitively high costs to produce large datasets.
Solution
Deployed Encord’s micro-model module to increase clinician labelling efficiency and automate 97% of produced labels.
Results
Highest expense clinician saw 16x labelling efficiency improvement. Cut model development time from 1 year to 2 months.
6.4X
Faster than manual labelling
97%
Automated labels
6X
Faster to AI in production
Memorial Sloan Kettering Cancer Center
MSK adopted Encord to build custom label structures for pulmonary thrombosis projects.
Problem
Detecting and classifying vena cava filters in complex label structures (ontologies) rendered existing & open-source tools unusable.
Solution
Deployed Encord’s ontology studio to build custom ontologies, DICOM annotation tool, & automation modules to increase efficiency.
Results
Project made feasible by the flexibility offered by Encord's ontology study.
1000+
Feature types
10 minutes
Total setup time
100%
Auditable