Betaface API
Betaface API is a face detection and face recognition web service. It can scan uploaded image files or image URLs, find faces and analyze them. API also provides verification (faces comparison) and identification (faces search) services, as well able to maintain multiple user-defined recognition databases (namespaces).
Faces general info:
- multiple faces detection (positions, sizes, angles)
- 123 face landmarks locations (22 basic, 101 pro)
- cropped face images
Click on the image to see all face landmarks you can get from the API:
Classification:
- estimate gender, age, ethnicity, emotion (smile/neutral)
- detect glasses, mustache and beard
Extended measurements:
- face and facial features description (shape, relative size and location)
- eyes, hair, skin, clothes and background colors
- facial hair amount
- approximate hairstyle (length, thickness, form)
Following functions are supported:
- upload image file or submit image URL.
- retrieve image and face metadata, including cropped face image.
- compare single faces or groups of faces and receive similarity confidence along with match decision.
- transform face image(s) – generate averages from two or more faces, change face expression or otherwise modify them.
- add user defined metadata tags, store user-adjusted points and face info.
Our API let you maintain and perform fast searches in very large faces collections (1 million faces or more), arranged in namespaces. We can provide additional technical assistance if you have to index large amount of data.
Feel free to use this shared webservice for commercial or non-commercial applications. Depending on the amount of requests you send to webservice your images may be queued and processed with delay.
For commercial users we offer additional options, such as running dedicated image processing servers, complete webservice or SDK licensing and custom development. For additional information please Contact us with your project description.
Current data storage policy: Source images are removed from cache shortly after processing. Faces that have no person/namespace assigned and corresponding image metadata are cleaned up after 10 days (face IDs and image IDs will be invalidated).