Case Study: Protex AI improves computer vision model accuracy with Voxel51 FiftyOne

A Voxel51 Case Study

Preview of the Protex AI Case Study

Protex AI improves model accuracy with Voxel51

Protex AI, a company that provides an AI-powered platform for monitoring worker health and safety, needed to improve the accuracy of its computer vision models. The models, which use object detection, classification, and pose estimation, analyze CCTV footage to autonomously capture unsafe events in industrial facilities. To address their challenge, Protex AI turned to Voxel51 and its product, FiftyOne.

Voxel51's FiftyOne was implemented as a vital component in Protex AI's pipeline. The solution was used to validate dataset annotations, intelligently subsample datasets to ensure they were balanced, and visualize model predictions for debugging and accuracy assessment. By using Voxel51's tools for dataset curation and model evaluation, Protex AI was able to improve the performance and reliability of its computer vision models.


View this case study…

Voxel51

52 Case Studies