Infilect (InfiViz)

Infilect (InfiViz) — Large-scale labelled retail shelf imagery and derived execution metrics. Merchandisers and field reps photograph shelves with phones/tablets; ~5 million images per month are processed across ~400,000 stores monthly in 18+ countries, and a syndicated audit product pools multi-brand observation across 2,000+ US stores. Each image is machine-parsed against a 250,000+ SKU/UPC reference database into SKU presence, facings, share of shelf, planogram compliance, on-shelf availability, price and promotion state, and display identification; two visits to the same store and categor…

Infilect (InfiViz) offers (Alternative, Sentiment, Fundamental) — ~5,000,000 images processed per month across ~400,000 stores monthly in 18+ countries (stated) — on the order of 60 million images per year, each carrying SKU/facing/price labels, atop a 250,000+ SKU/UPC reference database; the US syndicated audit layer covers 2,000+ stores.

Retail shelf object-detection and instance-segmentation model training at scale (a hard, long-tail, real-world CV dataset with millions of labelled images); OCR and key-value extraction models on shelf price tags; SKU identification and visual search across 250K+ products; planogram reconstruction models (image-to-layout synthesis); out-of-stock and availability classification; promotional-compliance and display-recognition classifiers; synthetic-data and augmentation generation for retail vision; and CPG market-share, distribution and pricing analytics as a syndicated alternative to manual audit panels

The data is with 9 years of history.

Coverage spans US; Application Software, Food Distributors; alternative, sentiment, fundamental; equities.