Product recognition: Datasets & Models
Product recognition uses image classification to identify which retail product appears in a photo, matching it against a known catalog of SKUs. It turns a shelf, cart, or checkout image into a product label.
Product recognition
This workflow applies image classification to recognize which retail product appears in a photo, matching it to a known SKU in your catalog. It accepts single-item or cropped product images as input and returns a predicted product class, and you can build or fine-tune it using a product recognition dataset for tasks like catalog tagging, self-checkout, and inventory verification.
Try it out
Powered by general detection model
Run on custom image
Drop an image here or click to upload
Why automate product recognition
Manually identifying products across thousands of SKUs is slow, error-prone, and does not scale to the volume of images retailers generate from cameras, mobile apps, and self-checkout stations. Barcode scanning fails when labels are obscured, damaged, or absent, and human catalog tagging cannot keep pace with rotating assortments. A classification model automates recognition directly from product appearance, reducing manual labeling effort and enabling real-time decisions at the shelf or register. This lets teams flag misplaced items, verify planograms, and speed up checkout without relying on a scannable code.
Key benefits
Automated product recognition delivers consistent, repeatable labeling that scales from hundreds to tens of thousands of SKUs and integrates with existing store cameras, checkout hardware, and mobile capture. A ViT-based classifier captures fine-grained visual differences between similar packages and brand variants, improving accuracy on lookalike products. As the catalog changes, the model can be retrained on new images, keeping recognition current without rebuilding infrastructure.
How to use it
- Upload an image: Drop any photo into the widget above to try product recognition for free, right on this page.
- Preview instant results: The model detects cereal box, soda can, shampoo bottle, snack bag, canned food within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
- Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train Classification (ViT) on your own images for production accuracy.
- Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.
Recommended model for product recognition
When you're ready to build a production model for product recognition, we recommend training Classification (ViT). Unlike the ready-to-use model in the demo above, you fine-tune it on your own labeled images for higher accuracy, then deploy to a Serverless Cloud API endpoint.
Not sure where to start?
Describe your Product Recognition project to Roboflow Agent.