|
|
Perceptual learning and Ideal Bayesian obsever in visual search task
Němeček, Viktor ; Děchtěrenko, Filip (advisor) ; Pilát, Martin (referee)
Searching for objects in a complex environment is an activity we do many times each day. Najemnik and Geisler (2005; 2008; 2009) showed in their work that people do not perform optimally, and devised multiple ideal observer models for one particular visual search task. In this thesis we tried to show that if people get feedback from one of the ideal observer models, they learn to solve the task better during a given amount of trials than they would without the feedback. We were unable to prove any nontrivial result with statistical significance due to a small sample size, but the data suggests that the feedback indeed has a positive effect on the learning, and that the continuation of the research is justified. An iOS application necessary for the experiment was created as a part of the thesis. Aside from the experiment itself, one can also use it to play a visual search testing game. 1
|
|
|
Ideal Bayesian Observer with reduced detectability map
Amemori, Josef ; Děchtěrenko, Filip (advisor) ; Brom, Cyril (referee)
Title: Ideal Bayesian Observer with reduced detectability map Author: Josef Amemori Department: Department of Software and Computer Science Education Supervisor: Mgr. Filip Děchtěrenko, Department of Software and Computer Science Education Abstract: A computational modeling of the human vision is a challenging task. In recent years, a biologically inspired model Ideal Bayesian Observer was created for the visual search task (Najemnik & Geisler, 2005). The model predicts eye movements when searching for Gabor patch in 1/f noise. In their work, they observed similarity between distributions of fixations and saccades predicted by Ideal Bayesian Observer and distributions of fixations and saccades from a human observer. In this work, we have implemented Ideal Bayesian Observer with degenerated visual field and compared the model with behavior of a human. Keywords: Ideal Bayesian Observer, eye movements, modeling, central scotoma
|
guest ::
