Introducing SemanticMD Cloud
At SemanticMD we believe medical image analysis should be easy. Key to that is providing imaging scientists and researchers the tools to reduce data wrangling (see Data Janitor) and to scale machine learning algorithms beyond the typical limits of a script running on a single server. Our latest API release makes it easier than ever to scale your medical image analysis algorithms whether you are working on computer-aided diagnosis or detection. But creating datasets needed for machine learning can be tricky, so our developers at SemanticMD have been hard at work building the next step in our journey to make medical image analysis easier: SemanticMD Annotate, an on-demand cloud service to annotate medical images for training algorithms available today in private beta.
SemanticMD Annotate
Whether you’re dealing with DICOM data, ultrasound images, clinical videos, or even large Aperio format pathology images, scaling out to both annotate and analyze terabytes of data becomes trivial with SemanticMD Cloud. We’ll cover some prime SemanticMD use cases in later posts, but first let’s take a tour of what you’ll get with SemanticMD Annotate, our medical image annotation platform.
Want to build a quick tool to triage patients or prioritize worklists? With SemanticMD Annotate you can upload a set of images and have clinical experts quickly annotate the data using a friendly, web-based quiz.
Building a computer-aided detection algorithm or just want to be able to automatically highlight regions of interest in your images? With SemanticMD Annotate you can minimize physician annoyance by providing them an easy-to-use quiz where they can highlight ROIs in under a second and automatically train state-of-the-art object detection algorithms.
Point and click model creation
If a Unix terminal isn’t your preference, we still got you covered. The SemanticMD Dashboard gives you a simple place to train and test your image analysis algorithms — with no coding required. Within the dashboard you’ll be able to quickly determine the accuracy of your image analysis apps and get insight into how you can improve it.
When do you need SemanticMD Cloud?
SemanticMD Cloud solves a broad range of use cases ranging from basic medical image classification to tracking complex objects in videos through terabytes of data. SemanticMD Cloud in particular focuses on and excels at those times when you have complex, imaging data that needs to be annotated, analyzed and scaled out.
The usual solution is to spend hours cleaning and organizing datasets, annotating, and writing custom Matlab scripts. Even after this process, you are still left with the last mile problem of refining the algorithm until the accuracy is acceptable and finally throwing the script “over the fence” for someone to make it production ready. Our team at Carnegie Mellon University has experience developing hundreds of image analysis algorithms the hard way.
With SemanticMD, medical image analysis is simple: annotate data, use the API to train a classifier or object detection algorithm, and deploy. This removes your ceiling and lets you focus on the innovative aspects of medical image analysis. And of course, if you would rather have a managed service with SemanticMD Cloud and focus on your medical device or web application, our developers could help you with that.
Features and availability
SemanticMD Cloud is available today in private beta. The introductory plan includes a total of 1 TB of storage space for your medical imaging data and a G2 Double Extra Large compute node with 8 GPU cores (see Amazon EC2 Instances). Several other plans are available during private beta to make it easier for you to meet your needs.
All SemanticMD Cloud plans come with:
- Unlimited annotation projects
- Deep learning API (bonus: object recognition included)
- Modality-specific image preprocessing
- Hosted demo sites
- 1 TB of image/video storage
- 24/7 Support
Get started
Head over to SemanticMD to request access today. If you have questions feel free to drop us a line at info@semantic.md.
