Final Report Summary - FAST FILTERING (Fast Filtering for Computer Graphics, Vision and Computational Sciences)
The Fast Filtering research project has developed new computational tools and methods in computer vision and graphics with an emphasis on efficiency and minimal user input. One resulting method allows the computer to automatically detect and count a repeating object in an image. This is very useful for scientific research in biology and medicine (automated cell counting), as well as for industrial purposes (manufacturing lines inspection). A second method solves the problem of objects which are non-printable with current 3D-printers due to the overhangs structures they contain. The new algorithm performs extensive searches for such structures and breaks down the object into a minimal number of sub-parts each of which can be printed with no additional support material. The use of the fast filtering schemes developed allow this extensive search phase to be carried out in short running-times.