2021 International Conference on Network Communication and Information Security (ICNCIS2021)
Home
Prof. Donatella Giuliani

Prof. Donatella Giuliani

微信截图_20210909094402.png

Prof. Donatella Giuliani

University of Bologna,  Italy


Speech title:Eye Detection applying a Probabilistic Algorithm inspired to the Butterfly Flight


Abstract:

Eye state analysis is relevant to detect fatigue and drowsiness when driving, thus it is essential to ensure safety of people. In this work, we propose an eye detection method applied to color image, recurring to a random exploration performed simultaneously by multiple research agents. The color image is represented in YCbCr color space, because in this space the luminance component is separated from the chrominance components. Firstly, an algorithm is used to recognize the face area in the original image. Afterwards, a normalized EyeMap is generated to distinguish and enhance the eye regions in order to make them more evident. The search of ocular areas is then conducted by simulating the flight of butterflies. Initially a butterfly gives rise to a random research but, when a foraging zone is discovered, it makes non-random dispersal movements around the food source. Similarly, in the applied algorithm, if a small region containing eyes is detected - even if partially - a more circumscribed research gets started through helicoidal paths around the identified area. In this way, we avoid the arbitrariness of choosing the size and position of the search window. To conclude, we have carried out the analysis of eye motion between different frames using the optical flow, based on Lucas-Kanade method.


Brief biography:

Graduated in Physics at University of Bologna (Italy), in 1982 and attended a Master in “Theory and Applications of Computational Machines” at University of Bologna in 1983. From 1984 to 1988 she worked as Software Analyst for Geoseismic Data at Aquater Spa, a company of ENI Group. From 1988 to 1992 she was employed as Software Analyst in Cartographic and Photogrammetric Data at Ecobit Spa (Italy). Since 1992 she is professor of Mathematics of MIUR. She graduated in Mathematics (2000) at University of Camerino (MC), developing research activities on shape analysis and models of biological structures using non-Euclidean geometry. In 2001 she received a Master degree in “Methodologies of e-learning” at University of Florence, Italy (2001). Since 2003 she worked as a teacher at University of Bologna. In 2009 she received a Philosophical Doctorate in Mathematics and Statistics in Computational Sciences (MaSSC) at University of Milano, with research applications on computational methods and morphological analysis in Neuroimaging in cooperation with IRCCS Institute Fatebenefratelli (Brescia). In 2015 at University of Bologna, she received a Master degree as trainer for teaching of Mathematics