AI based technology to leverage the potential of face biometric data.
Our face recognition technology incorporates 3 models to reach the highest performance:
1) Sertis' SoTA face detection model to detect faces in an image
2) Sertis' SoTA face landmark model to accurately localize face landmarks (e.g. pupils, nose peak, mouth corners, etc.) and align the detected face
3) Our NIST certified face embedding model to map individual's facial features
The embeddings thus generated can be used for:
Validate if two images contain the same person based on their facial data. Examples include passport photos to webcam verification.
Look up the detected face against a database containing tens of thousands of faces to find an accurate match.
Rank #1 in Thailand and top 30 worldwide
On the open evaluation by NIST, our face recognition pipeline was certified as
Face Mask Detection
Accurately detect the presence of a mask covering a person's face. Our mask detection framework can be used by organizations and businesses to help monitor sanitary measures compliance.
Our framework contains two main modules:
Face detection and alignment
using Sertis' SoTA face detection and landmark models.
Face mask classification
to accurately predict the presence of a mask covering the detected face.
Combining Sertis' SoTA face recognition technology with body temperature measurement to allow entry into specified areas. With security as a priority, we incorporate our Presentation Attack Detection (PAD) model to detect and prevent presentation attacks (e.g. printed and video replay attacks) from fooling our biometric system.
Precisely identifies customers remotely with our digitalized and electronic process of KYC equipped with Sertis' SoTA face recognition technology and our PAD model to effectively prevent falsification of identity and both passive and active presentation attacks.