Case Study: Volpara Solutions reduces mammography errors with Microsoft Azure Stream Analytics

Leading health provider revolutionizes mammography, reduces errors with custom computer vision models

Volpara Solutions, a New Zealand–based health provider focused on mammography analytics, faced the challenge of identifying poor-quality or non-standard images without relying on unreliable metadata. This was especially important for excluding implant and quality control phantom images from analysis so clinicians could avoid errors and reduce unnecessary recalls. The company used Microsoft Azure Stream Analytics alongside Azure Cognitive Services, including Custom Vision, to address the issue.

Microsoft Azure Stream Analytics helped Volpara Solutions build custom image classification models that analyze pixel data directly, allowing clinical staff to reclassify images and retrain models through an easy-to-use interface. The result was a faster, more reliable way to detect images that should be excluded from statistics, improving mammography quality control and supporting fewer analysis errors. Microsoft’s solution enabled model training in minutes and gave Volpara the flexibility to deploy the models in Azure or export them for local use.

Volpara Solutions