Abstract

An investigation of the relative speed and effectiveness of two computer vision algorithms has been conducted. One algorithm incorporates a two-level data hierarchy. The other incorporates a one-level hierarchy and serves as a relatively conventional basis for comparison. The computer vision algorithms, programmed in Fortran, detect and recognize a moving square. Both computer vision algorithms could readily be implemented in existing hardware. The two-level algorithm was found to be up to 90% faster than the one-level algorithm. An analysis was made of elapsed CPU time variance as a function of time of day and user load. This was done to minimize the variance of results in comparing the above two algorithms. The mean and standard deviation of elapsed CPU time were both found to increase with system load, and system load was found to exhibit a midday peak.

Publication Date

4-10-1986

Document Type

Thesis

Student Type

Undergraduate

Degree Name

Imaging Science (BS)

Department, Program, or Center

School of Photographic Arts and Sciences (CIAS)

Advisor

Willem Brouwer

Comments

Physical copy available from RIT's Wallace Library at TA1632 .Z64 1986

Campus

RIT – Main Campus

Download

Share

COinS