Improving the efficiency of labour-intensive visual effects tasks is the aim of this industry collaboration project which is midway through its 18-month trajectory.
AutoPost or ‘Deformable Surface Tracking and Alpha Matting for the Automation of Post Production Workflows’ is a project funded by the EC under H2020 which started in January 2015 and runs until June 2016.
Project partners (Eurecat, Fraunhofer HHI, imcube labs, Moto and Digital Graphics) are at the IBC Future Zone presenting the objectives and expected results of AutoPost compared with commercially available software tools.
This includes a breakdown of visual effects where matting and tracking are commonly used and a demonstration of the limitations of existing workflows and tools.
“We are also presenting a new dataset which has been created to evaluate and compare matting and tracking techniques under varying real-world conditions,” explainedEurecat’s R&D technical project manager, Monica Caballero.
The aim of AutoPost is to improve the efficiency of labour-intensive vfx tasks by integrating “state-of-the-art” algorithms for natural video matting and deformable surface tracking into plug-ins for established post production platforms.
The matting and tracking plug-ins developed within the AutoPost project can be used in 2D and 3D post production workflows for a variety of vfx including object manipulation, scene extension and appearance modification.
“We are especially targeting small and medium-sized facilities which may not have the resources to either develop their own in-house solutions or use very specialised and expensive green screen, motion capture or rotomation techniques,” she said.
Autopost Project. 8.F13