GET IN TOUCH
OUR INNOVATIVE TECHNOLOGY
Combining our state-of-the-art face segmentation and single view 3D face reconstruction, we achieve face swapping in quality and scale never seen before. Our face recognition technology is the first to provide a bidirectional mapping, both to the face embedding space and back to pixel space. Our face style transfer technology can transform faces to cartoons or different animals and back. Our main focus currently is to extend our technology to work on videos, which will include both detection and tracking of faces.
IMAGE-TO-IMAGE FACE SWAPPING
OUR FACE SWAP
Previous face swapping techniques will most likely fail for non-frontal faces and in conditions that are not perfect. Our face swap was designed to work in extreme conditions: different face poses, lighting conditions, image quality, face shape, and skin color. We can handle occlusions of common objects like hair, glasses, and hands. Tested on millions of images, we can guarantee our method is the most robust face swap available.
WHY FACE SWAP?
Face swapping is one of the most important techniques of automatic face image editing. It has numerous practical applications such as hairstyle replacement, privacy protection, and data augmentation for machine learning. Most importantly it's funny and viral.
We’re a talented group of entrepreneurs and engineers experienced in Deep Learning and developing web-based applications. Using the power of deep convolutional networks, we solve practical problems in visual understanding and make our solutions available to the public for free. At DEEPNEN, we believe that the main disadvantage of Deep Learning is that it's not accessible enough for non experts, we strive to change that. We provide web-applications for both businesses and independent users in various fields, mainly in face synthesization.