NEURALSEG: A STATE-OF-THE-ART MACHINE LEARNING PLATFORM USING DEEP LEARNING TO SEGMENT MRI IMAGES
NEURALSEG: A STATE-OF-THE-ART MACHINE LEARNING PLATFORM USING DEEP LEARNING TO SEGMENT MRI IMAGES
Challenge
NeuralSeg is a state-of-the-art machine learning platform that automatically segments MRI images with the power of deep learning. NeuralSeg reduces the time required for MRI image segmentation from days of manual work to a few minutes of machine work and the cost from thousands of dollars to a fraction. At its core, NeuralSeg’s algorithm uses GPU processing to perform the segmentation.
Challenge
NeuralSeg is a state-of-the-art machine learning platform that automatically segments MRI images with the power of deep learning. NeuralSeg reduces the time required for MRI image segmentation from days of manual work to a few minutes of machine work and the cost from thousands of dollars to a fraction. At its core, NeuralSeg’s algorithm uses GPU processing to perform the segmentation.
How
we helped
How
we helped
Cloud & DevOps Implementation
UI/UX Development
Roadmap Planning & Strategy
Grant Funding Support
Team Building & Knowledge Transfer
Cloud & DevOps Implementation
UI/UX Development
Roadmap Planning & Strategy
Grant Funding Support
Solution
NeuralSeg’s machine learning algorithm was developed by researchers at McMaster University. The team engaged NuBinary to turn their scientific algorithm into a commercialized web-enabled and scalable platform deployed on cloud and easily accessible by their clients. NuBinary designed and implemented NeuralSeg platform’s web interface, user and image management system, heavy file uploading system, online MRI image preview, async backend job processing, job dispatching and scaling system. We also distributed cloud infrastructure for best cost efficiency, reliability and scalability.
Solution
NeuralSeg’s machine learning algorithm was developed by researchers at McMaster University. The team engaged NuBinary to turn their scientific algorithm into a commercialized web-enabled and scalable platform deployed on cloud and easily accessible by their clients. NuBinary designed and implemented NeuralSeg platform’s web interface, user and image management system, heavy file uploading system, online MRI image preview, async backend job processing, job dispatching and scaling system. We also distributed cloud infrastructure for best cost efficiency, reliability and scalability.
Technology
Stack
- Ruby on Rails and Google Cloud PostgreSQL on the backend
- React.js front-end
- Google Cloud file storage for large file upload support
- Google Cloud compute instances and distributed Cloud GPU instances for processing parallelization
- Docker containers
- Continuous integration and delivery using CircleCI
- Near-full automated test coverage on the backend and the front-end for highest quality
Technology
Stack
- PHP Laravel
- React.js front-end
- Elasticsearch Big data storage and query engine
- RabbitMQ for message caching and dispatching
- Continuous Integration and Deployment
- Auto scaling, load balanced and fault tolerant on DigitalOcean nodes
TELL US ABOUT YOUR PROJECT. LET’S TAKE YOUR BUSINESS WHERE IT NEEDS TO BE.
Tell us about your project, let’s take your business where it needs to be.