Proceedings: GI 2008

Evaluation of techniques for visualizing mathematical expression recognition results

Joseph LaViola, Anamary Leal, Timothy Miller, Robert Zeleznik

Proceedings of Graphics Interface 2008: Windsor, Ontario, Canada, 28 - 30 May 2008, 131-138

  • BibTex

    author = {LaViola, Joseph and Leal, Anamary and Miller, Timothy and Zeleznik, Robert},
    title = {Evaluation of techniques for visualizing mathematical expression recognition results},
    booktitle = {Proceedings of Graphics Interface 2008},
    series = {GI 2008},
    year = {2008},
    issn = {0713-5424},
    isbn = {978-1-56881-423-0},
    location = {Windsor, Ontario, Canada},
    pages = {131--138},
    numpages = {8},
    publisher = {Canadian Human-Computer Communications Society},
    address = {Toronto, Ontario, Canada},


We present an experimental study that evaluates four different techniques for visualizing the machine interpretation of handwritten mathematics. Typeset in Place puts a printed form of the recognized expression in the same location as the handwritten mathematics. Adjusted Ink replaces what was written with scaled-to-fit, cleaned up handwritten characters using an ink font. The Large Offset technique scales a recognized printed form to be just as wide as the handwritten input, and places it below the handwritten mathematical expression. The Small Offset technique is similar to Large Offset but the printed form is set to be a fixed size which is generally small compared to the written expression. Our experiment explores how effective each technique is with assisting users in identifying and correcting recognition mistakes with different types and quantities of mathematical expressions. Our evaluation is based on task completion time and a comprehensive post-questionnaire used to solicit reactions on each technique. The results of our study indicate that, although each technique has advantages and disadvantages depending on the complexity of the handwritten mathematics, subjects took significantly longer to complete the recognition task with Typeset in Place and generally preferred Adjusted Ink or Small Offset.