Case Study: reverse.fashion Scales Textile Sorting AI with Voxel51

A Voxel51 Case Study

Preview of the Reverse.fashion Case Study

reverse.fashion saves hundreds of engineering hours with Voxel51

reverse.fashion, a company building AI for textile sorting and recycling in Europe, faced a challenge with its five-person machine learning team managing millions of ultra-high-resolution garment images. The core problem was not a lack of data but the difficulty of effectively curating it—deciding which samples to label and ensuring the labels accurately reflected the complex, subjective real-world task of sorting garments. They turned to the vendor Voxel51 to avoid building their own infrastructure and to focus on model development.

The solution involved implementing Voxel51 as a shared visual layer for their ML workflow. This platform provided synchronized access for the entire team, enabling visual data inspection, integration with labeling tools, and model evaluation without cumbersome handoffs. As a result, reverse.fashion saved hundreds of engineering hours previously spent on manually preparing data for review. The Voxel51 solution also allowed domain experts to contribute directly, improving label quality and ensuring their models solved the correct practical problem.


View this case study…

Voxel51

52 Case Studies