Case Study: Altris achieves 12% higher model accuracy with SuperAnnotate

How Altris AI achieved 12% increase in model accuracy by switching to SuperAnnotate

Altris AI, an ophthalmology AI company using computer vision and deep learning to support retinal disease diagnosis on OCT scans, needed a better way to manage in-house annotators and improve the quality of segmentation and classification work. They had been using Labelbox for segmentation and AWS SageMaker for classification, but struggled with poor annotation usability, low-quality pathological area labels, and inefficient workflows for their medical doctor annotators.

SuperAnnotate provided the platform Altris AI switched to for annotation management, offering a more robust toolset, easier user experience, and stronger project management and analytics. With SuperAnnotate, Altris AI increased model accuracy from 80.1% to 92.4%, reduced annotation time from 5.8 to 4.9 minutes per scan, and cut QA time from 2 to 1 minute per scan, while also lowering annotation costs and improving overall annotation quality.

Altris

Andrey Kuropyatnyk

Chief Executive Officer