Towards the automated identification of Chrysomya blow flies from wing images. Issue 3 (15th April 2018)

• Record Type: Journal Article
• Authors: Macleod, N.; Hall, M. J. R.; Wardhana, A. H.
• Is Part Of: Medical and veterinary entomology. Volume 32:Issue 3(2018)
• Page Start: 323
• Page End: 333
• Publication Date: 2018-04-15
• Subjects: Chrysomya bezziana -- Chrysomya megacephala -- Chrysomya rufifacies -- Calliphoridae -- automated identification -- biogeography -- image classification -- morphometrics -- Old World screwworm fly -- sexual dimorphism; Entomology -- Periodicals; Veterinary entomology -- Periodicals; Insects as carriers of disease -- Periodicals; 616.968
• DOI: 10.1111/mve.12302
• Languages: English
• ISSNs: 0269-283X

Abstract

Abstract: The Old World screwworm fly (OWSF), Chrysomya bezziana (Diptera: Calliphoridae), is an important agent of traumatic myiasis and, as such, a major human and animal health problem. In the implementation of OWSF control operations, it is important to determine the geographical origins of such disease‐causing species in order to establish whether they derive from endemic or invading populations. Gross morphological and molecular studies have demonstrated the existence of two distinct lineages of this species, one African and the other Asian. Wing morphometry is known to be of substantial assistance in identifying the geographical origin of individuals because it provides diagnostic markers that complement molecular diagnostics. However, placement of the landmarks used in traditional geometric morphometric analysis can be time‐consuming and subject to error caused by operator subjectivity. Here we report results of an image‐based approach to geometric morphometric analysis for delivering wing‐based identifications. Our results indicate that this approach can produce identifications that are practically indistinguishable from more traditional landmark‐based results. In addition, we demonstrate that the direct analysis of digital wing images can be used to discriminate between three Chrysomya species of veterinary and forensic importance and between C. bezziana genders. Abstract : Direct morphometric analysis of Chrysomya wing images can be used to discriminate among species, populations and genders with high levels of accuracy. Graphic summarization of morphological differences based on image‐derived discriminant functions can be of substantial value in the interpretation of morphological variation. These developments suggest that fully automated, flexible, user‐friendly identification systems for Chrysomya (and other insect groups) that would be of substantial medical and veterinary value are practically achievable.