Determining Points on Handwritten Mathematical Symbols
This addresses a practical bottleneck in processing handwritten mathematics and diagrams for applications like OCR and digital handwriting enhancement.
The paper tackles the problem of automatically identifying key feature points (like baselines) in handwritten mathematical symbols of varying sizes and irregular layouts, enabling applications such as improved 2D mathematical recognition and handwriting neatening while preserving style.
In a variety of applications, such as handwritten mathematics and diagram labelling, it is common to have symbols of many different sizes in use and for the writing not to follow simple baselines. In order to understand the scale and relative positioning of individual characters, it is necessary to identify the location of certain expected features. These are typically identified by particular points in the symbols, for example, the baseline of a lower case "p" would be identified by the lowest part of the bowl, ignoring the descender. We investigate how to find these special points automatically so they may be used in a number of problems, such as improving two-dimensional mathematical recognition and in handwriting neatening, while preserving the original style.