The Next Generation of AI in Gastroenterology

The Next Generation of AI in Gastroenterology

Full stack AI development platform powered by endoscopy data

Full stack AI development platform powered by endoscopy data

Full stack AI development platform powered by endoscopy data

EndoML enables researchers at health systems, pharma, CRO, and med device companies to build precision medicine models.

EndoML enables researchers at health systems, pharma, CRO, and med device companies to build precision medicine models.

The EndoML Value

The EndoML Value

What can EndoML do for you?

What can EndoML do for you?

Develop and Deploy AI Models for Endoscopy

Develop and Deploy AI Models for Endoscopy

Researchers of varying skill levels can go directly from data to trained model, ready for deployment.

Researchers of varying skill levels can go directly from data to trained model, ready for deployment.

Discover Predictive Biomarkers

Discover Predictive Biomarkers

EndoDINO provides powerful features that can unlock novel insights from your clinical trial data.

EndoDINO provides powerful features that can unlock novel insights from your clinical trial data.

Secure & Compliant

Secure & Compliant

Built with security at its core, EndoML is both HIPAA and SOC 2 compliant, deployed securely within your cloud environment.

Built with security at its core, EndoML is both HIPAA and SOC 2 compliant, deployed securely within your cloud environment.

Powered by EndoDINO

Powered by EndoDINO

The new gold standard backbone model for AI development in endoscopy.

The new gold standard backbone model for AI development in endoscopy.

EndoDINO:
A New Standard in AI for GI

EndoDINO:
A New Standard in AI for GI

EndoDINO is an AI foundation model for GI endoscopy, trained on Virgo's industry-leading dataset. The model achieves SOTA performance in a variety of GI endoscopy benchmarks and generalizes well to new data. EndoDINO offers unmatched performance and accelerates research and development.

EndoDINO is an AI foundation model for GI endoscopy, trained on Virgo's industry-leading dataset. The model achieves SOTA performance in a variety of GI endoscopy benchmarks and generalizes well to new data. EndoDINO offers unmatched performance and accelerates research and development.

Key Differences from Etro-FM & Endo-FM

Key Differences from
Etro-FM & Endo-FM

Etro-FM

5,145

Endo-FM

33,000

Endo-FM

33,000

EndoDINO (ours)

130,037 (<10% of available)

Etro-FM

Initial Videos

Endo-FM

Etro-FM

Removal of low quality

EndoDINO
(ours)

Frame Curation

Initial Videos

Endo-FM

Varying frame rates

5,145

EndoDINOv1 (ours)

Hierarchical K-means clustering

33,000

Etro-FM

525,711

130,037

Frames Used

Frame Curation

Endo-FM

5,024,101

Removal of low quality

EndoDINOv1 (ours)

10,000,000

Varying frame rates

Etro-FM

DINOv1

Hierarchical K-means clustering

Training Architecture

Frames Used

Endo-FM

Spatial-temporal

525,711

EndoDINOv1 (ours)

DINOv2

5,024,101

10,000,000

Training Architecture

DINOv1

Spatial-temporal

DINOv2

Features

Features

Object Detection and Segmentation

Object Detection and Segmentation

EndoML supports the development of object detection and segmentation models. Rapidly annotate your data and train video segmentation models. EndoDINO provides powerful features for all sorts of anatomy and pathology, ranging from polyps to ampullae, and everything in between.

EndoML supports the development of object detection and segmentation models. Rapidly annotate your data and train video segmentation models. EndoDINO provides powerful features for all sorts of anatomy and pathology, ranging from polyps to ampullae, and everything in between.

EndoML supports the development of object detection and segmentation models. Rapidly annotate your data and train video segmentation models. EndoDINO provides powerful features for all sorts of anatomy and pathology, ranging from polyps to ampullae, and everything in between.

Natural Language Video Analysis

Natural Language Video Analysis

Analyze your endoscopy videos with the power of natural language. This enables users of all skill levels to unlock valuable insights seamlessly. Researchers, clinicians, and pharmaceutical teams can securely upload datasets and use the chat interface to ask complex questions.

Analyze your endoscopy videos with the power of natural language. This enables users of all skill levels to unlock valuable insights seamlessly. Researchers, clinicians, and pharmaceutical teams can securely upload datasets and use the chat interface to ask complex questions.

Insightful Visualizations

Insightful Visualizations

Interactive tools like T-SNE and Sankey diagrams provide clear, data-driven visualizations. These tools help users uncover patterns and insights to better understand and utilize endoscopy data with the ability to visualize insights in different ways based on the unique features extracted by EndoDINO.

Interactive tools like T-SNE and Sankey diagrams provide clear, data-driven visualizations. These tools help users uncover patterns and insights to better understand and utilize endoscopy data with the ability to visualize insights in different ways based on the unique features extracted by EndoDINO.

API Access for Power Users

API Access for Power Users

Our dynamic API access ensures seamless integration with existing systems. Featuring “Beast Mode” for power users, EndoML empowers developers with customizable tools to unlock the full potential of their data.

Our dynamic API access ensures seamless integration with existing systems. Featuring “Beast Mode” for power users, EndoML empowers developers with customizable tools to unlock the full potential of their data.

Data Exchange

Data Exchange

EndoML's Data Exchange feature allows companies to license anonymized, high-quality endoscopic video datasets to train their own AI models. This unlocks opportunities for researchers and pharmaceutical companies to enhance their machine learning capabilities with real-world data.

EndoML's Data Exchange feature allows companies to license anonymized, high-quality endoscopic video datasets to train their own AI models. This unlocks opportunities for researchers and pharmaceutical companies to enhance their machine learning capabilities with real-world data.

How it Works

How it Works

Our foundational model is powered by EndoDino, the most powerful model specifically trained on endoscopic videos.

Our foundational model is powered by EndoDino, the most powerful model specifically trained on endoscopic videos.

Securely upload your endoscopy videos to the EndoML platform.

EndoDINO analyzes the videos and provides immediate insights.

Access advanced tools to run experiments and fine-tune models.

Deploy models for research and clinical care.

Ready to Transform Your Research?

Ready to Transform Your Research?

Discover how EndoML can revolutionize your approach to gastroenterology AI. Contact us today or schedule a demo to see the power of EndoDINO in action!

Contact Us

The Next Generation of AI in Gastroenterology

The Next Generation of AI in Gastroenterology

© 2024 All rights reserved.