Case Study: Trendspek optimizes structural crack detection with Datature

Trendspek Optimizes Structural Crack Detection with Datature

Trendspek, a provider of 3D asset management software, needed to enhance its inspection process by augmenting it with AI to quickly identify structural defects like cracks in concrete. They partnered with the MLOps vendor Datature to develop a proof-of-concept machine learning model without needing to write code, leveraging the Datature Nexus platform to build a solution that could integrate with their existing system.

Using Datature's end-to-end platform, Trendspek efficiently annotated data and trained a semantic segmentation model within 24 hours. The solution, deployed via Datature's Inference API, enabled Trendspek to automatically detect cracks and render them on their 3D models, speeding up the defect identification process by up to five times. This integration of Datature's technology allows Trendspek to offer users faster, AI-powered insights while maintaining crucial human oversight for quality assurance.

Trendspek

Josh Sinclair

Head of Product