Case Study: PlantSnap achieves scalable, high-accuracy plant identification (320K species, 90% top-5 precision) with Imagga's Custom Categorization API

Training the world’s largest plant recognition classifier

PlantSnap, a mobile app company that helps users identify flowers, trees, mushrooms and other flora worldwide, faced a massive scalability and accuracy challenge: there are over 320,000 plant species to classify and many look‑alike species that confound training. PlantSnap turned to Imagga and its cloud-based Custom Categorization API to handle the large-scale image training required (using a dataset of over 90 million images) and to deliver an enterprise-grade image recognition service for iOS and Android.

Imagga implemented its Custom Categorization solution on powerful NVIDIA DGX Station hardware, enabling PlantSnap to train models 10x faster, resolve look‑alike species issues, and scale to 320K categories. The result: Imagga-powered PlantSnap achieved a 90% precision rate for the top 5 returned results, launched its mobile app within a week (saving over six months of development), and could focus on user acquisition while relying on Imagga for hosting and model deployment.

PlantSnap

Eric Ralls

Founder & CEO