An algorithmic method for the recognition of fungal spore cells in microscopic images, as well as its development and its origin, are described and demonstrated. The process is designed for a machine vision project which automatically identifies fungal spores within field samples for epidemiological simulation models. The method consists of a three-pass system that successfully recognizes spores in any position and which is tolerant of occlusion. / The algorithm, as implemented, demonstrated an accuracy of $ pm$5.3% on low quality images which is less than the assumed error of humans performing the same task. The processing speed also compared favorably with the performance of humans. / The method developed presents a framework of description that, through the first two passes, highlights certain distinctive aspects within an image. Those highlighted aspects are then recognized by the third pass. The system is loosely based on biological vision, is extremely versatile and could be adapted for the recognition of virtually any object in a digitized image.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.27282 |
| Date | January 1997 |
| Creators | Bernier, Thomas. |
| Contributors | Landry, J. A. (advisor) |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Master of Science (Department of Agricultural and Biosystems Engineering.) |
| Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
| Relation | alephsysno: 001578381, proquestno: MQ29656, Theses scanned by UMI/ProQuest. |
Page generated in 0.0017 seconds