Andre Barczak

Doctor of Philosophy, (Computer Science)
Study Completed: 2007
College of Sciences

Citation

Thesis Title
Feature-based Rapid Object Detection: From Feature Extraction to Parallelisation

Read article at Massey Research Online: MRO icon

Mr Barczak studied rapid object detection in computer vision. His work focused on overcoming some of the limitations of current methods used when objects are subject to geometric transformations. Among the contributions of his work, a new feature extraction method was developed. This method extends the usability of moment invariants because it increases the dimensionality of the feature set by extracting features from concentric areas of the image. Advantages of the method include rapid computation and invariance to rotation, scaling and lighting conditions. His work can be applied to areas such as video processing, surveillance, optical character recognition, and image-based quality assurance.

Supervisors
Dr Martin Johnson
Dr Chris Messom