We use Deep Learning to drive industry-leading Face Analysis and Generation

Based on a proprietary new class of deep-learning neural networks optimized to exploit geometry for speed, scale and accuracy, combined with a unique large-scale high resolution 4D dataset, Facesoft’s models have been proven in international competition to be among the best in the world. They can be used for multiple applications including facial detection, recognition, generation, expression, dense 3D motion capture and more. The Facesoft team has multiple patents pending and has published over 200 papers in the field.

Industry Leaders in Face Analysis Technologies

Face detection
Face Detection

State of the Art face detection and analysis

Face landmarks
Face Landmarks

Robust, fast landmark localisation algorithm

Face matching
Face Matching

State of the Art face verification technology

Face recognition
Face Recognition

Fast face recognition, for millions of identities

Face generation
Face Generation

3d avatars from still images or videos

Facesoft is a company formed by a world-renowned group in machine learning and computer vision for face analysis

tick
Cutting edge machine learning algorithms with a proprietary large-scale 4D dataset
tick
We harness the power of Generative Adversarial Networks (GANs) for generating million-scale faces
tick
Technology to generate realistic 3D avatars from single images and videos
tick
Proprietary technology for training Deep Convolutional Neural Networks (DCNNs)
tick
State-of-art technology in face recognition, detection, generation and emotion analysis
tick
We capitalize on the power of geometric deep learning for generating and analyzing 3D faces

Working in Collaboration with

Imperial College London logo