A Modular Web-Based Platform for Structured Data Collection and Annotation
Jakub Tomeš, Daniela Janstová, Jan Mareš · 2025
Efficient acquisition, management, and annotation of data are critical in research involving structured datasets. Conventional methods, such as local file storage, often lead to fragmented data, limited accessibility, and workflow inefficiencies. To address these issues, we developed a modular web-based platform integrating a recording module, annotation interface, and centralized database to support structured data handling and accessibility. The platform was initially developed for hand gesture recording but is designed to be data-type agnostic. It includes integrated labeling, real-time synchronization via Firebase, and support for VR headsets. Both cloud and self-hosted deployment options are available, allowing flexible storage and security configurations. Current development extends the platform to Raman spectral data management, focusing on spectrum inspection rather than conventional labeling. Implemented tools include baseline correction, filtering, normalization, and export functions to support dataset preparation. The platform aims to simplify data handling and improve the efficiency of research workflows across various scientific fields. This work presents a versatile, open-ended platform enabling researchers to capture, annotate, and manage structured data in real-time, with minimal technical overhead and maximum flexibility.