Vascular Segmentation App

ML-driven medical app automates vascular analysis in 3D-printed tissue. Uses advanced ML models, OpenCV preprocessing, graph algorithms, and 3D structure generation for precise evaluations.

Client
Advanced Solutions Life Sciences
Date
June 18, 2020
Service
Machine Learning/Software Engineering
Location
Louisville - KY (Remote)
View live project
Advanced Solutions Life Sciences

Challenge

This project focuses on developing a machine learning-powered medical application for automated evaluation of vascular features in synthetic 3D-printed tissue and bone. It employs advanced ML models and microscopy projection samples, utilizing both Desktop and Cloud platforms.

Challenges:

  1. Image Preprocessing: Implementing efficient OpenCV preprocessing methods for enhancing image quality and accuracy.
  2. Algorithm Optimization: Designing and integrating graph-based algorithms for precise inference post-processing.
  3. Backend Development: Leading the creation and deployment of a serverless backend using Innovation Lab Fabric for seamless operations.
  4. External Runtime: Developing a robust external runtime in Python for accurate inference execution.
  5. 3D Structure Generation: Creating an algorithm to construct a detailed 3D representation of vascular structures based on inference results.

‍

Advanced Solutions Life Sciences

Project Results

Successfully overcoming the challenges, the project has established a seamless, automated pipeline for identifying vascular structures and features in synthetic 3D-printed tissue and bone. Through efficient OpenCV preprocessing, image samples were optimized, enhancing accuracy. Graph-based algorithms were meticulously implemented, ensuring precise inference post-processing. The development and deployment of a serverless backend using Innovation Lab Fabric streamlined operations, while the external runtime in Python enabled accurate and swift inference execution. The culmination of these efforts resulted in the creation of a sophisticated algorithm capable of constructing intricate 3D representations of vascular structures. The customer now possesses a cutting-edge solution, empowering them with an advanced tool for the automated evaluation of vascular features, marking a significant milestone in medical science and technology.