Foundry (formerly “The Foundry”) has found a new CEO to replace Alex Mahon. Mahon recently succeeded David Abraham as chief executive-designate of Channel 4. Mahon remains on Foundry’s board, as does Bill Collis who she replaced less than two years ago. Mahon is now non-executive deputy chairman of the company.
Craig Rodgerson, formerly CEO of Capita Digital Software and Services, takes up the role effective immediately. “We are thrilled that Craig has joined Foundry,” said Nic Humphries, senior partner of London-based HgCapital, the private equity firm behind Foundry.
“His excellent track record of driving sustainable growth within the software space is the perfect fit to support Foundry’s ongoing strategic direction.” Rodgerson is seen as somewhat of a return to the company’s engineering and research roots, as he has a very strong technical engineering background, and he has an excellent history of delivering innovative software and meeting shipping targets. In other words, he is known for running tight cooperative software teams that ship product.
In recent days Foundry has started discussing some of it newer moves into areas such as Deep Learning technologies in Nuke.
Facilitating these new moves are strategic partnerships with UK and European research institutes and universities. Two of the internships the company has been running this year have been exploring the impact of AI on video editing, image processing, understanding and manipulation. “We believe AI and deep learning could have a massive impact on visual effects – as it is already doing so in other industries – and these research projects are allowing us to dip our toes into the water of this exciting field” commented the company with the release of the details of two specific projects:
Project #1: Traditional Effects The first project explores the potential of deep learning to remove noise from images and videos in post-production. Decades of research and handcrafted algorithms have already gone into code to clean up images through signal process analysis.
PhD Student Da Chen, from Bath University, joined the NukeX team at Foundry to test whether deep learning could be used to remove noise in the Nuke compositing software. “Deep learning has enabled us to take hundreds of thousands of examples of noisy input images – and images without the noise – and train a neural network how to map from the noisy image to a clean image on the output, quickly and efficiently” remarked the company in their release. Rather than painstakingly handcrafting the noise reduction process, Foundry is teaching NUKE to produce results automatically.
Project #2: Seeing the Light The second project explored harnessing AI to ‘hallucinate’ physical lighting from just a single picture so that digital models can be added photorealistically. On set, the real world lighting is normally captured with HDR probes or bracketed image sets. The second project is exploring estimating lighting IBL approaches by training an AI system to reproduce a plausible 3D lighting setup from a simple input image. This is very important in AR. Say an iPad is looking down at a desk, it is important to estimate the 360 lighting so the 3D placed in the scene matches the background lighting.
PhD student Corneliu Ilisescu, from University College London, joined Foundry’s cinematic Augmented Reality (AR) team. The team has explored how deep learning can be used to recognize lighting using thousands of examples where the lighting is known, and training the network to set up lighting for AR content. This project has explored how AI can make light capture (estimation) simpler and more accessible. It does not replace light probes, rather it learns from a world of light probes and can estimate a plausible solution very quickly with reduced input.