Face Detection API
The face detection endpoint locates every face in an image and returns structured data for each one. Use it to count people in a frame, crop face regions for downstream processing, or quality-gate images before recognition or embedding extraction.
Data Returned per Face
- Location — Bounding box (x, y, width, height) for the face region.
- Detection confidence — Model certainty score; filter on this to drop low-quality detections.
- Facial landmarks — Pixel coordinates for eyes, nose tip, and mouth corners; use for alignment before embedding.
- Face size — Pixel dimensions; filter out small or distant faces that will degrade recognition accuracy.
- Attributes — Predicted gender (use with care; treat as a signal, not ground truth).
Common Use Cases
- People counting in a frame or scene.
- Cropping individual face regions before embedding extraction or emotion analysis.
- Quality gating — reject images where no face is found or confidence is below threshold before charging for recognition.
Priced at $0.001 per image in production. Free sandbox available.
Face Embeddings | API Pricing | Developer API overview