Detecting Behavioural Motifs of Mosquitoes Using Deep Variational Embedding of Posture Dynamics
Keywords
Loading...
Authors
Issue Date
2023-10-01
Language
en
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Mosquitoes are considered to be one of the most dangerous animals as they
have been responsible for countless number of deaths in the past. Even to this day,
mosquito-borne diseases such as malaria and dengue continue to claim millions
of lives in humid regions around the world. Recent advancements in the study
of mosquito behavior have shed light on alterations in the behaviour of infected
mosquitoes which facilitates the easier spread of diseases. The quantification of
these behavioral alterations offers a promising avenue for the reduction in infection
transmission. Investigating the genetic influence on behaviour could help establish
correlations between specific genes and their impact on mosquito biting behaviors,
potentially leading to the development of disease-resistant ”super mosquitoes” that
cannot transmit infections to other animals.
In the field of biology and medicine, deep learning approaches have helped
scientists explore new avenues. From tumor detection to wildlife tracking, biologists
can now explore novel theories and approaches. The combination of different
machine learning methodologies to address complex problems has consistently
yielded reliable results. Drawing inspiration from the success of deep learning,
our work introduces a comprehensive pipeline using techniques such as pose estimation,
tracking, dimensionality reduction, and clustering to quantify observed
behavioural alterations in mosquitoes.
This project introduces a methodology, wherein we utilize mosquito videos to
recognise and analyze their behaviours. We compare the behaviors exhibited by
a population of dengue-infected mosquitoes with those of an uninfected control
population. Our primary focus lies in the identification of behaviours most crucial
for disease transmission and ascertaining whether the distinctions between the
control and dengue-infected mosquito populations align with recent findings in the
field.
Description
Citation
Faculty
Faculteit der Sociale Wetenschappen