Akademik Veri Yönetim Sistemi
2014 22nd Signal Processing and Communications Applications Conference, SIU 2014, Trabzon, Türkiye, 23 - 25 Nisan 2014, ss.108-111, (Tam Metin Bildiri)
In this study, is aimed to estimated of the region which observed on the screen by the artificial neural network (ANN) trained with using images of laptop user. Study were formed of two phases. In the first stage, an ANN has been developed for determining eye region in the captured image with webcam. In the second stage, the data obtained from the eye region are used for training of another ANN which will be estimated to region of coordinate. For the creation of the training set, the laptop screen is divided to 57×32 pieces region of coordinate (24×24 pixels in size). 100 pieces of these regions identified as random. And user when looking at each of this regions were taken 20 different images via internal webcam. For the creation of test set randomly selected 20 regions. And each of this regions were taken 20 different images. In the training set was used 2000 pieces image. Similarly, test set was created using 400 images. The result of the test set processing with ANN, the region of coordinates in close proximity to the actual region of coordinates were obtained. Of the regions estimated via the ANN, was seen to cluster around of actual region and does not interfere the values obtained for the different zones. © 2014 IEEE.