Case Study: Perceptly speeds up ad annotation and boosts accuracy with Labellerr

How Perceptly Uses Image Similarity To Fasten Up Annotation

Perceptly helps ad agencies analyze their ad creatives to determine what elements drive performance. Their challenge was manually tagging 50-60 elements per ad from a large volume of images and videos, a process that was too slow and inefficient for their analytical engine. This manual bottleneck hindered their ability to generate insights quickly for campaign planning.

Labellerr's automated training data generation platform addressed this by using image similarity to copy relevant tags between files, saving over 70% of the time compared to manual work. This solution, which also included features for quick review and consensus-based tagging, achieved more than 95% accuracy in data extraction. As a result, Perceptly significantly streamlined its analytical process and enhanced the accuracy of its ad performance insights.

Perceptly