A research team from Radboud University in the Netherlands has used deep learning techniques to convert sketches facial images into actual photographs.Using a new method called Convolutional Sketch Inversion, the team has reversed the process seen in popular apps whereby faces are changed into sketches.The innovation could have important implications for forensics, in turning sketches based on eyewitness accounts into something a photo-recognition tool could use, said the team.To train and test the algorithm, the scientists used computer-generated sketches based on the photos in the CelebA dataset, an online resource with more than 200,000 celebrity images, and in the LFW dataset, a collection of 13,000 images of faces from the web. The hand-drawn sketches were taken from the CUFS dataset.The researchers, who work in the university's Donders Institute for Brain, Cognition and Behaviour are now looking to see how they can bring their work to market. They are looking at applications in the fine arts and forensic arts world for monetizing their work.
Recent Video
- New non-profit trade association of HAND will standardise global talent ID
- No cards, no phones: J.P. Morgan’s biometric revolution in payments
- Anthony Vidiano on building the future of access management
- Charisma Check wins 2025 Start-Up Pitching Competition, advancing identity verification and background checks for dating app users
- How to protect one billion Amazon customers: Lessons from continuous risk assessments
