The Neglected Role of GUI in Performance Evaluation of AI-Based Transcription Tools for Handwritten Documents

This paper aims to inspect the often neglected role of Graphical User Interfaces (GUI) in AI-based tools designed to assist in the transcription of handwritten documents. While the precision and recall of the handwritten word recognition have traditionally been the primary focus, we argue that the time parameter associated with the GUI, specifically in terms of validation and correction, plays an equally crucial role. By investigating the influence of GUI design on the validation and correction aspects of transcription we want to highlight how the time that the user must take to interact with the interface must be taken into account to evaluate the performance of the transcription process. Through comprehensive analysis and experimentation, we illustrate the profound impact that GUI design can have on the overall efficiency of transcription tools. We demonstrate how the time saved through the utilization of an assistant tool is heavily dependent on the operations performed within the interface and the diverse features it offers. By recognizing GUI design as an essential component of transcription tools, we can unlock their full potential and significantly improve their effectiveness.