Digital-Square-Tanzania/rdt-capture

RDT Capture # Table of Contents * Overview * Installation * Configuration for New RDTs * API * Attribution * About Us * License # Overview RDTScan is an open-source library for developers who are interested in creating Android apps that support the digital curation and interpretation of rapid diagnostic tests (RDTs). RDTScan provides the following functionality: **1. Real-Time Quality Checking During Image Capture** RDTScan uses image processing to check the quality of images intercepted from the smartphone's camera while the user moves their smartphone over the RDT. RDTScan provides functions for checking the blurriness and lighting of incoming camera frames. If the image is in the camera's field-of-view, RDTScan also checks the scale and orientation of the RDT in the image. To help end-users capture the clearest image possible, RDTScan intelligently generates instructions based on these quality checks. **2. Robust Result Interpretation** Assuming a suitable image has been captured, RDTScan can post-process the image to emphasize any faint lines that may appear on the immunoassay. The end-user can view the post-processed image for themselves to make a more informed decision about the RDT results. Alternatively, RDTScan provides a lightweight algorithm that interprets the test results on the end-user's behalf. RDTScan uses a feature-matching approach for RDT recognition, allowing RDTScan to be quickly adapted to new RDT designs. Unlike model-driven approaches that require a dataset of example images for model training, RDTScan only requires a single example image and some additional metadata (e.g., relative position and meaning of each line) to detect and interpret RDTs. RDTScan is intentially designed so that all of the algorithms can run on the smartphone, which means that there is no need to upload confidential photos to a server to use this library. **Disclaimers:** * Although RDTScan has been tested through multiple in-lab studies and real-world deployments, this libr …