|M.Sc Student||Koren Yelena|
|Subject||Motility Phenotyping of Model Organism C. Elegans Using|
Scale-Invariant Feature Transform (SIFT)
|Department||Department of Biomedical Engineering||Supervisor||Professor Josue Sznitman|
|Full Thesis text|
Caenorhabditis elegans (C. elegans) is perhaps the best understood metazoan in terms of anatomy, genetics, development, and behavior. This small, 1mm long worm is widely used to investigate fundamental questions in biology, in particular for behavioral genetics, as well as for drug screening and development and for modeling different aspects of human diseases.
Studying the locomotory behavior of C. elegans is frequently used to uncover genetic traits. Such approaches are helpful for instance in understanding the basis of genetic diseases in humans and devising future treatment strategies.
Characterizing nematode motility phenotyping is typically based on the classification of nematodes according to their movements. Since traditional approaches to classify behavioral patterns of C. elegans locomotion have been usually based on manual detection performed by human experts, phenotyping has often been imprecise and qualitative in nature. As a result, algorithms for automated classification of C. elegans motility phenotypes have been developed over the years to characterize locomotion in a more quantitative fashion. Current modern computer vision techniques for motility phenotyping rely on defining and extracting physical metrics such as amplitude, wavelength, frequency, speed and curvature amongst other, in an effort to quantify locomotion. Here, rather, we propose the combination of Scale-Invariant Feature Transform (SIFT) and Bag-of-Words image representation, a widely-used statistical method in computer vision, to provide an automated approach for analyzing motility behavior relations between different strains of C. elegans.