For event photography teams

AI photo matching for event photographers.

PolyReID brings event uploads, face match, OCR, AI-enhanced visual organization, focused human review, gallery publishing and a branded storefront into one web workspace.

No credit card required to explore and test the workflow.

App, model and API

Four layers with different responsibilities.

AI photo matching describes a set of retrieval signals. Face match, OCR and visual matching can prepare candidates, while the app connects those signals to an operational event-photo workflow.

The application

The PolyReID web app keeps event uploads, person groups, review states, kiosk access and the branded storefront in one workspace.

The appearance model

Face match, OCR and visual models produce signals that help prepare candidate groups for review. They are suggestions, not automatic identity decisions.

The customer experience

Configure event names, covers, themes and group labels, then share a kiosk QR code or a direct store link with customers.

The developer API

The Person ReID API is a narrower building block: one authorized person crop returns a versioned appearance embedding, without the surrounding photo workflow.

One visible workflow

From event upload to reviewed gallery.

Each stage has a bounded job. AI matching prepares suggestions; a person reviews ambiguity before customer-facing publication.

01

Upload the event

Bring a live or completed high-volume photo shoot into one event workspace.

02

Match the useful signals

Use face match, OCR, visual matching and event context to prepare candidate photo groups.

03

Review the suggestions

Confirm, reject, merge or split proposed groups while keeping unresolved cases visible.

04

Publish reviewed work

Move approved event galleries into a branded storefront without treating raw suggestions as final.

Human review by design

Keep the reviewer between a suggestion and a published gallery.

Similar clothing, occlusion, motion blur and partial views can create plausible but incorrect visual matches. Reviewers can confirm, reject, merge, split or leave a group unresolved before publication.

Appearance similarity is not proof of identity.

The app prepares candidate groups from visual signals. It does not assign a name or make an identity determination; the customer-facing gallery follows reviewed state, not raw model output.

Test before you commit

Test AI photo matching on your next event.

Start with the event-photography workspace and explore face match, visual organization and human review without a credit card.