Object classification methods for application in FPGA based vehicle video detector
Year: 2009
Type: article
Abstract: The paper presents a discussion of properties of object classification methods utilized in processing video streams from a camera. Methods based on feature extraction, model fitting and invariant determination are evaluated. Petri nets are used for modelling the processing flow. Data objects and transitions are defined which are suitable for efficient implementation in FPGA circuits. Processing characteristics and problems of the implementations are shown. An invariant based method is assessed as most suitable for application in a vehicle video detector.
Source: Transport Problems
Author: W. Pamuła
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1537571226