Fig 1. Overview of the data collection and the developed model for transcribing BP images.

Upper left: The screenshots of the mobile apps employed in the Guatemala Perinatal, Georgia I3 and Georgia IMPROVE studies are shown. Upper right: In the Guatemala Perinatal study, images were initially stored locally before being uploaded to cloud storage. However, in the Georgia studies, data uploading was done directly from mobile devices. Lower left: Datasets utilized for model training, validation, and testing, including Guatemala Set 1, Guatemala Set 2, Georgia I3, and Georgia IMPROVE are identified. The studies employed an Omron M7 automated oscillometric BP monitor for the Guatemala datasets, and Omron 3 devices for the Georgia datasets. Lower right: The BP image transcription model consists of two steps: 1) LCD Localization and 2) Digit recognition.