We wanted to let the community know about a new image database (with associated landmark & semi-landmark data for the images) that we have published. Briefly, this database contains both the raw digital images & associated (semi)landmark data for Drosophila wings across multiple controlled genotypes and sexes. For each individual we have imaged both left and right wings, and for each wing on two different microscopes and two different magnifications. For all of this data we have also acquired the full set of landmarks and semilandmarks using the "WINGMACHINE" pipeline developed from the lab of David Houle.
In the associated paper (describing all of this) we have also done some basic comparisons of machine learning (and simple) classification approaches (on genotype and sex) for the landmark data. This was compared to using computer vision + machine learning approaches (i.e. biometric approaches) that utilize very different features than we normally use in geometric morphometric analysis. The goal of the database was to be helpful to the community and we have tried to provide everything (raw data, scripts, extensive meta-data on provenance of samples). So we hope this will be of use to you. The paper can be found here http://www.gigasciencejournal.com/content/4/1/25 The database can be found here (including all images, data and scripts) http://dx.doi.org/10.5524/100141 and links to just scripts and landmark data are also available on the lab github page https://github.com/DworkinLab/Wing_Biometrics_2015 -- Ian Dworkin Department of Biology McMaster University -- MORPHMET may be accessed via its webpage at http://www.morphometrics.org To unsubscribe from this group and stop receiving emails from it, send an email to morphmet+unsubscr...@morphometrics.org.
