What scenarios does face-searching support? 
Common use cases include access control, visitor management, security surveillance, attendance, member recognition, blacklist alerting, and public area search.
How do I integrate face-searching?
Integration is usually via API or SDK:
- Apply for service and obtain API keys
- Create a face database
- Add face images to the database
- Call the search API for matching
- Receive similarity scores and results
How many faces can one database store?
Capacity varies by plan, ranging from thousands to millions of faces.
How fast is face retrieval?
Single-face search is usually millisecond-level, depending on database size and hardware.
What similarity score indicates a match?
Scores range from 0 to 1. A threshold of 0.7–0.9 is recommended; high-security scenarios may use 0.85 or higher.
Does it work with masks, glasses, makeup, or low light?
The algorithm is robust to mild occlusion and lighting changes, but heavy occlusion, extreme blurriness, or strong backlighting will reduce accuracy.
Can it mistake twins or lookalikes?
There is a small risk of false matches. For high security, combine face recognition with ID verification.
