Insect identification using Artificial Neural Networks (ANN)
| Název česky | Determinace hmyzu pomocí ANN |
|---|---|
| Autoři | |
| Rok publikování | 2008 |
| Druh | Konferenční abstrakty |
| Fakulta / Pracoviště MU | |
| Citace | |
| Popis | Introduction: Entomology, as well as its application in many fields, such as agriculture,forestry, human and veterinary medicine, relies heavily on the accurate identification of species. Besides molecular diagnostic techniques, the progress in information technology has opened opportunities for the computer-assisted taxonomy. Artificial Neural Networks (ANN) seem to have been one of the most promising tools for the basis of such systems. The advantages of ANN include an ability to learn from examples and to generalize observed patterns. Methods: The use of ANN requires a training database in which specimens, correctly identified by experts, are included. Each specimen has to be characterized by diagnostic variables (characters). For ANN inputs can be used digital images, optically sensed wing beat frequency spectra, near-infrared reflectance spectra, bioacoustic recordings, chemotaxonomy or morphometry. An ANN model is designed to find a relationship between the characters (=input) and species (=output). The quality of the training set is an essential prerequisite to obtaining reliable identifications. Results: Our case studies on thrips and diptera used morphometric data mostly. The high percentage of correctly identified specimens (about 97 %) is promising for a wider use of ANN for insect identification in practice. Conclusions: ANN is cheap and non-destructive suitable also for type material or permanently mounted slides. ANN have the potential to enhance the practice of routine identification with a non-expert as technical help. High reliability of classification is promising for a wider application of ANN in the practice of insect identification. |
| Související projekty: |