Contact

Endoscopy

Powering your next innovation in AI for endoscopy

Challenges in medical video are varied. Soft tissues deform, cameras jitter, occlusions arise, and even the experts may disagree with one another in interpretation. RSIP Vision has tackled these challenges and time and again helped our clients take their products to the next level.

AI and Computer Vision Technologies

Our broad suite of endoscopic image analysis capabilities includes image segmentation, 2D-to-3D reconstruction in multiple modalities, tissue classification, motion estimation, and more.

Abnormality & Tissue classification

Detect & distinguish subtle findings in endoscopic video and MRI

+
Abnormality & Tissue classification
Endoscopy with cancer detection
  • Automatic detection of tumors, ulcers, other pathologies in gastroscopy videos
  • Automated analysis of MRI and contrast CT scans
  • Accurate classification of findings
  • Save time and effort for treating physician

3D Reconstruction

Reconstruct anatomy and implements from 2D imaging

+
3D Reconstruction
Applications
  • Lumen depth reconstruction
  • Scope / catheter reconstruction
  • Tool reconstruction
Input types
  • Stereo or monocular video
  • Fluoro & other intraoperative imaging modalities
Technologies
  • Neural networks - when training data is available
  • Classical computer vision

Image Segmentation

Segment & measure findings and regions

+
Image Segmentation
  • Pixel wise and instance segmentation of anatomy and abnormalities
  • Challenges addressed:
    • Clinical variability between patients
    • Image & color variability between scopes
    • Correct handling of low-quality frames
    • Ambiguous & deforming soft tissue anatomy
    • Subtle visual differences between distinct anatomical classes
Video-Analytics

Video Analytics

Detect key events & abbreviate videos

+
Video Analytics
  • Video-AnalyticsIdentify workflow stages
  • Detect operations of implement (e.g. biopsy collection)
  • Entering & exiting various parts of the anatomy
  • Detect occlusion of the field due to blood, wash, motion, etc
  • Detect proximity to anatomical structures
  • Video abbreviation - auto-detection of non informative frames
EUS

Endoscopic Ultrasound

Identify views and landmarks. Image registration and reconstruction

+
Endoscopic Ultrasound
  • EUSLabeling of anatomical areas of interest
  • Precise segmentation
  • 3D model creation from 2D cross sections
  • Automated image quality grading
  • AI based artifact detection
  • Multimodal registration

Advanced Endoscopy & ERCP

3D reconstruction, stricture classification, tool tracking, and more

+
Advanced Endoscopy & ERCP
  • 3D reconstruction from fluoro shots
  • Improve navigation using computer vision
  • Detection of anatomical areas and events in endoscopic video feed
  • MRCP anatomical analysis
  • Tracking tools and implements
  • Classifcation of strictures

Multimodal Registration

Fuse multiple image sources

+
Multimodal Registration
  • Classical & deep learning based technologies
  • Customized to user needs
  • 2D-to-2D, 2D-to-3D, and 3D-to-3D capabilities
  • Rigid and deformable registration
  • Support all modalities & image characteristics

Motion Estimation

Track scope, objects, and anatomy

+
Motion Estimation
  • Leveraging deep learning and classical image processing techniques
  • Tracking tissue & objects in field of view
  • Tracking self-motion of scope
  • Applicable to challenging surgical, endoscopic, and endoluminal scenes

Customized R&D

Application and device-specific solutions, from research through full product development

Click here

Case study

Endoscopic Video Analysis

The challenge
  • An innovative start-up initiative in the endoscopy space wanted to quickly prove out a range of AI benefits for video analysis application
Our approach
  • The RSIP Vision R&D team, in close collaboration with in-house and client medical experts, developed a multi-functional proof of concept
  • We quickly & efficiently built key capabilities including pathology detection, image segmentation, motion estimation, and image quality improvement
  • During the development process, RSIP Vision discovered novel AI use cases with strong additional business potential, enhancing our client’s value proposition
The outcome
  • We supported our client, leading to a successful transaction with a large industry-leading corporation
  • We developed & delivered novel core technologies, providing a strong basis for the development of a full solution 
  • Production and solution development is underway in continued partnership with RSIP Vision

Learn more

Learn more solutions and technologies from this field

Next Generation Intra-op Navigation

During minimally invasive surgeries, surgeons are limited in their field of view, which restricts their observation of the patient’s full anatomy. This may lead to suboptimal surgical outcomes or even errors. For example, in a tumor excision procedure, healthy tissue may be injured, or worse, the surgeon may fail to optimally remove the lesion. Surgeons base the procedure on pre-op images of the area such as CT scans, which provide a snapshot of the internal tissue. Navigation systems are then utilized to fit the scan to the coordinate system of

Read More
AI in Diagnostic ERCP

Image Analysis and AI in Diagnostic ERCP

Recent developments in the field of medical image analysis and artificial intelligence (AI) are used to improve the procedural outcomes of ERCP (Endoscopic retrograde cholangiopancreatography). Here is how RSIP Vision develops AI for diagnostic ERCP. Read what we do for enabling 3D Image Reconstruction and Image Registration/Fusion. Learn how strictures detection and classification can provide the physicians with classification scoring and, sometimes, help them avoid unnecessary biopsies during ERCP.

Read More
Type 4 cholangiocarcinoma

Enhanced ERCP tumor assessment using AI

ERCP involves both endoscopy and fluoroscopy on a region with limited access. Accordingly, it poses several challenges for the gastroenterologist performing the procedure: imaging and artificial intelligence are key technologies to solve these challenges. They are used to reconstruct multiangle 2D X-ray images into a 3D image that will help the gastroenterologists in real-time navigation, reducing any complications involved in the navigation process. AI can be trained to accurately classify each tumor found by ERCP and check whether it is cancerous or not. 

Read More
ERCP Gallstones and strictures

AI in ERCP gallstones and strictures treatment

When performing ERCP for stone removal and stricture treatment, the gastroenterologist must overcome several challenges, namely navigation in the region of interest and the choice of the most fitting treatment that must be selected and implemented. AI enables to reconstructing a 3D image from multi-angle X-ray images or ultrasound slices. It is also trained to accurately classify the type of blockage that necessitated the ERCP procedure in the first place, resulting in quicker and more efficient ERCP procedures.

Read More

Gastrointestinal lesion detection with machine learning

Endoscopic examination is performed to disclose the lesion’s biophysical properties and assess its severity according to its physical appearance. Computer vision and machine learning put a large arsenal of techniques at our disposal. Clever utilization of these techniques enables to do just that, that is, translating expert knowledge into a fine-tuned algorithm, specifically designed for the task of GI lesion detection. The highest level of robustness, accuracy, and reproducibility is required, hence automatic methods can perform as a proper alert system for lesion detection. 

Read More

Next Generation Intra-op Navigation

During minimally invasive surgeries, surgeons are limited in their field of view, which restricts their observation of the patient’s full anatomy. This may lead to suboptimal surgical outcomes or even errors. For example, in a tumor excision procedure, healthy tissue may be injured, or worse, the surgeon may fail to optimally remove the lesion. Surgeons base the procedure on pre-op images of the area such as CT scans, which provide a snapshot of the internal tissue. Navigation systems are then utilized to fit the scan to the coordinate system of

Read More
AI in Diagnostic ERCP

Image Analysis and AI in Diagnostic ERCP

Recent developments in the field of medical image analysis and artificial intelligence (AI) are used to improve the procedural outcomes of ERCP (Endoscopic retrograde cholangiopancreatography). Here is how RSIP Vision develops AI for diagnostic ERCP. Read what we do for enabling 3D Image Reconstruction and Image Registration/Fusion. Learn how strictures detection and classification can provide the physicians with classification scoring and, sometimes, help them avoid unnecessary biopsies during ERCP.

Read More
Type 4 cholangiocarcinoma

Enhanced ERCP tumor assessment using AI

ERCP involves both endoscopy and fluoroscopy on a region with limited access. Accordingly, it poses several challenges for the gastroenterologist performing the procedure: imaging and artificial intelligence are key technologies to solve these challenges. They are used to reconstruct multiangle 2D X-ray images into a 3D image that will help the gastroenterologists in real-time navigation, reducing any complications involved in the navigation process. AI can be trained to accurately classify each tumor found by ERCP and check whether it is cancerous or not. 

Read More
ERCP Gallstones and strictures

AI in ERCP gallstones and strictures treatment

When performing ERCP for stone removal and stricture treatment, the gastroenterologist must overcome several challenges, namely navigation in the region of interest and the choice of the most fitting treatment that must be selected and implemented. AI enables to reconstructing a 3D image from multi-angle X-ray images or ultrasound slices. It is also trained to accurately classify the type of blockage that necessitated the ERCP procedure in the first place, resulting in quicker and more efficient ERCP procedures.

Read More

Gastrointestinal lesion detection with machine learning

Endoscopic examination is performed to disclose the lesion’s biophysical properties and assess its severity according to its physical appearance. Computer vision and machine learning put a large arsenal of techniques at our disposal. Clever utilization of these techniques enables to do just that, that is, translating expert knowledge into a fine-tuned algorithm, specifically designed for the task of GI lesion detection. The highest level of robustness, accuracy, and reproducibility is required, hence automatic methods can perform as a proper alert system for lesion detection. 

Read More
Show more

Get in touch

Please fill the following form and our experts will be happy to reply to you soon

Recent News
All news
Upcoming Events
Stay informed for our next events
Subscribe to Our Magazines

Subscribe now and receive the Computer Vision News Magazine every month to your mailbox

 
Subscribe for free
Follow us
Linkedin Twitter Facebook Youtube

contact@rsipvision.com

Terms of Use

Privacy Policy

© All rights reserved to RSIP Vision 2021

Created by Shmulik

R&D Services

How we work

Planning
    • Free consultations for concept development
    • Work plan creation & literature review
Data
    • Data acquisition assistance thru our medical partners
    • Support custom datasets

Execution

    • Multidisciplinary team – PM, R&D, medical, & annotation
    • Weekly client meetings
    • Full transparency at all stages

Deliverables

    • Software, source code, technology transfer, integration support & new IP
    • Solution is yours to keep, no per-use royalties