Tencent Cloud FaceID leverages AI technologies such as document OCR recognition, liveness detection, and one-to-one face comparison to verify individual identity. It enables you to verify a user's identity in seconds, improving your service efficiency and reducing labor costs.
FaceID offers OCR, liveness detection, and one-to-one face comparison services to meet your specific needs.
1. Compare with data from authoritative databases
2. Compare with an uploaded photo
The OCR feature recognizes both front and back sides of an ID card in just one scan and returns all fields, including ID number, name, and validity period. Its number recognition accuracy exceeds 99.9%, placing it first in multi-field recognition. It also performs well in some special situations such as skew, dark light, exposure, greatly improving the recognition accuracy.
Liveness detection can meet the high security requirements of facial recognition in different verification scenarios. It can analyze real-time videos to decide whether they come from a live person or a fake, and guard against presentation attacks such as photos, videos, and static 3D models. Currently, it supports three liveness detection modes: lip, action, and silent.
Face comparison can automatically verify the identity of a person by calculating the similarity of their liveness face and library face image based on facial features. It leverages the large-scale face datasets collected by Tencent's social networking platforms, while tapping into technologies such as high-dimensional LBP, PCA, LDA and Bayesian networks, metric learning, transfer learning, and deep learning. With an accuracy rate of 99.80% as tested during LFW in 2017, it supports face detection and recognition in multiple scenarios regardless of age, posture, or lighting conditions.