Biologically plausible approach is an alternative to conventional engineering approaches when developing algorithms for intelligent systems. It is apparent that biologically inspired algorithms may yield more expensive calculations when comparing its run time to the more commonly used engineering algorithms. However, biologically inspired approaches have great potential in generating better and more accurate outputs as healthy human brains. Therefore more and more new and exciting researches are being experimented everyday in hope to develop better models of our brain that can be utilized by the machines. This thesis work is an effort to design and implement a computational model of neurons from the visual cortex's MST area (medial superior temporal area). MST's primary responsibility is detecting self-motion from optic flow stimulus that are segmented from the visual input. The computational models are to be built with dual Gaussian functions and genetic algorithm as its principle training method, from the data collected through lab monkey's MST neurons. The resulting computational models can be used in further researches as part of motion detection mechanism by machine vision applications, which may prove to be an effective alternative motion detection algorithm in contrast to the conventional computer vision algorithms such as frame differencing. This thesis work will also explore the interaction effect that has been discovered from the newly gathered data, provided by University of Rochester Medical Center, Neurology Department.
Library of Congress Subject Headings
Computer vision; Intelligent control systems; Artificial intelligence; Machine learning; Genetic algorithms; Neurons--Computer simulation
Department, Program, or Center
Computer Science (GCCIS)
Yu, Chen-Ping, "Computational model of MST neuron receptive field and interaction effect for the perception of self-motion" (2008). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus