Optimization of Sequential Filters
1995 (English)In: Proceedings of the SSAB Symposium on Image Analysis: Linköping, Sweden, 1995, 87-90 p.Conference paper (Refereed)
A recursive method for separation of spherically separable quadrature filters into simple kernels with mainly one dimensional extent has been worked out. The resulting filter responses are mapped to a non biased tensor representation where the local tensor constitutes a robust estimate of both the shape and the orientation (velocity) of the neighbourhood. The performance of this General Sequential Filter concept has exceeded the authors most optimistic expectations. A qualitative evaluation results in no detectable loss in accuracy when compared to conventional FIR (Finite Impulse Response) filters but the computation is performed 20-30 times faster. The magnitude of the attained speed-up implies that complex spatio-temporal analysis can be performed using standard hardware, such as a powerful workstation, in close to real time. Due to the soft implementation of the convolver and the tree structure of the sequential filtering approach the processing is simple to optimize for most standard hardware. The method used in the examples was implemented in AVS (Application Visualization System) using modules written in C.
Place, publisher, year, edition, pages
1995. 87-90 p.
Engineering and Technology
IdentifiersURN: urn:nbn:se:liu:diva-21748OAI: oai:DiVA.org:liu-21748DiVA: diva2:273873