MONITORING MALARIA/DENGUE/ZIKA VECTORS AND MAPPING POSSIBLE SPOTS OF FUTURE OUTBREAK USING COMPUTER VISION AND MACHINE LEARNING

Abstract

Swarms of mosquitoes erupting from soil-logged eggs (exacerbated by climate change) carrying a host of diseases such as Zika, dengue and malaria, making mosquitoes the world's deadliest animals. Outbreaks of such vector borne diseases leaves drastic impact especially on subtropical countries with limited public health resources. Lack of competitive surveillance system and poorly developed diagnostic infrastructure are one of the reasons for poor understanding of distribution and magnitude of these diseases. Chances of outbreak of Zika virus (ZIKV) following human West Nile virus (WNV) and chikungunya (CHIKV) epidemics has left the public in panic. Unavailability of licensed vaccine to deal with such epidemics to date leaves vector control only preventive measure in such conditions. This Project aims to monitor such outbreaks by taking in account depending environmental variables and mapping possible spots that'll help allocate limited public health resources. The system comprises of mosquito surveillance system that consist of two parts one for classifying and counting mosquitoes in real time and other part provides mosquito observatory environment for behavioural analysis. Real time classification, trend and behavioural analysis is then visualized on Mosquitrack (Dashboard created using JavaScript, Python, Flask, ChartJS and leaflet respectively). This will help epidemiologists and health care sector in taking necessary steps to deal with such situation beforehand.