Profile Detail

Femiani, John

Assistant Professor
Department of Engineering

Office: PRLTA 230U
Phone: (480) 727-5108
FAX: (480) 727-1248
Email: john.femiani@asu.edu


My main research interests are in the area of computer vision and visualization, combining computer graphics and pattern recognition. I am very interested in the software side of converting physical artifacts into editable digital content. Major research areas are a) segmentation of raster images, point clouds, or surfaces, and b) raster to vector conversion and reverse engineering.

Segmentation: Segmentation is an important part of low level computer vision because it reduces a dense set of data points into a relatively sparse set of parts, much like tokenization on strings. The goal is find pieces which can be manipulated with fewer interactions, or interpreted with less computation than the original data. I am particularly interested in hierarchical or multi-scale segmentation approaches, and ways to define practical generic models that can be used as a basis for knowledge based segmentation.

Vectorization and Reverse Engineering: Vectorization is a process of fitting a set of geometric primitives, such as lines or curves, to data points in order to form a more concise representation. Raster to vector conversion has applications to computer aided design, geographic information systems, and graphic design. This is a hot topic which includes translating aerial photographs and satellite images into maps, and translating images of text into machine-editable text. Vectorization is also a way to compress images, resize images without pixilation, animate images, and manipulate objects in images rather than the pixels themselves. Vectorization is an important research area because telescopes, satellites, digital cameras and cell phones allow lay people and scientists to produce new digital images which need to be manipulated.