Figure 3 : Overview of client-server communication and image analysis.

From: High-throughput and automated diagnosis of antimicrobial resistance using a cost-effective cellphone-based micro-plate reader

Figure 3

Using our custom-developed mobile application, three images of the sample plate taken at different exposure times are uploaded to our local or remote server. Our automated algorithm finds the wells using the blue channel of the brightest exposure image and extracts the average well intensity. The well intensities for the three exposures are then combined to maximize the dynamic range of our measured intensities and normalized respective to the maximum transmittance as obtained from blank plate control wells. These well intensities are then thresholded based off a database of uploaded plates and processed to determine each drug’s minimum inhibitory concentration and susceptibility. These results are stored on the server along with the uploaded plate images, and also returned to the client application to be shown to the user.