Microorganism Detection using Single-Shot Multibox Detector

The advent of microorganism detection has proved invariably useful to the scientific community, especially considering the recent pandemics across human history. There is a continuous rise in the occurrence of pandemics across the world which has made the need for innovative and swift detection methodologies more important. The current study proposes the usage of Single-Shot Multibox Detector (SSD) for detecting microorganisms which is unexplored till date. The proposed model utilizes EMDS-6 Dataset which has 840 images spanning 21 distinct microbe classes, with 40 images per class. The chosen model, SSD MobileNet V2 FPNLite 320×320 from the TensorFlow 2 model zoo has demonstrated more than satisfying results for performance with good accuracy and precision in detecting microbes. The results show the robustness and versatility of this SSD-based approach. The discussion of the results is also done along with the future scope of the current project.