Case Study: Surgical Data Science Collective achieves 10x faster video annotation with Encord

How Surgical Data Science Collective (SDSC) Conducted Video Annotation 10x Faster

Surgical Data Science Collective (SDSC) faced significant challenges in annotating its vast amount of surgical procedure video data. Their previous tool had slow performance and resulted in a 20% annotation error rate, compounded by a lack of customizability and integration capabilities for its machine learning pipelines. This created a major bottleneck in their efforts to democratize access to surgical data.

To solve this, SDSC implemented Encord's Training Data Platform, utilizing its Annotate and Active products. The solution provided fast, native video rendering, automated review workflows, and a Python SDK for seamless pipeline integration. As a result, Encord enabled a 10x increase in annotation speed, moving the team from completing two procedures every two months to a projected twenty procedures in four months, while also drastically reducing errors.

Surgical Data Science Collective

Margaux Masson-Forsythe

Director of Machine Learning (ML)