Surveillance and diagnosis of new and emerging communicable diseases in remote regions, such as the Amazon, is a challenging task. These regions can be difficult to reach, are sparsely populated, and have limited medical and ICT infrastructure. Medical practitioners and community health agents who work in such regions often have very basic qualifications, and therefore have limited knowledge of new and emerging diseases. The increasing capabilities of mobile devices, such as tablets and smart phones, have made them a useful platform for delivery of medical services in remote locations. In this paper we introduce a system that could potentially support diagnosis of vector-borne diseases such as Bartonellosis and Leishmaniasis in areas where specialist healthcare is scarce. In particular, we focus on the image analysis and classification component of this system, which aims to reduce the chance of misdiagnosing these less common diseases as malaria.