BlogCases

Development of a Proof of Concept for an AI-based software application for a Medtech project

Cases

ExplORer Surgical focuses on real-time surgical procedure data acquisition and management. The ExplORer system is a workflow management tool designed to manage tasks and tools for surgical teams.

They help surgical teams coordinate their tasks and tools by providing interactive workflow management software for use during surgical operations.

Business Task

The Dysnix team signed a contract for the development of Proof of Concept for AI-based software application to recognize surgical instruments from the video stream. This application is to be a part of the interactive workflow management software for use during surgical operations. The developed demo became the basis for attracting investment to implement the project further - www.crunchbase.com/organization/explorer-surgical

The Dysnix team's tasks were:

  • Dataset preparation
  • Application development
  • QA and bug fixes within 2 milestones (see below)
  • Deployment of an application into the customer's environment

The demo was developed to minimize risks.

Non-completed tasks and incorrect tools lead to:

  • Disruptions, delays, and errors in the operating room
  • Frustrations for the surgical team
  • Excess costs for hospitals
  • Suboptimal outcomes for patients.

Solutions

We prepared a dataset and trained a model for surgical instrument detection, as well as developed an application using this model. The application can:

  • Detect all the surgical objects on a video
  • Classify all the detected objects
  • Track objects on the screen and produce events on changes

Value delivered by Dysnix

  • Preparation of a dataset and training of a model for surgical instrument detection, as well as the development of an application using this model
  • All the development stages were shown as demos weekly so that customers could see progress in the work made
  • Trained the in-house team on how to use the developed solutions
  • Preparation of the documentation for obtaining financing
Industry

MedTech

Location
Chicago, IL USA
Project duration
4 months (October 2019 - January 2020)
Our team
6 (1 CTO, 2 Computer Vision Engineers, 1 Full-Stack Developer, Senior DevOps Engineer, Project Manager)
Technologies used

Pytorch

OpenCV

MLflow

InceptionNet

Resnet

LabelMe

Numpy

Pandas

Pillow

Matplotlib

Scikit

Services provided