2026 ASEE Annual Conference & Exposition

Identifying, categorizing, and evaluating the taxonomy of STEM content and the obstacles in converting handwritten lecture materials of engineering courses for improved accessibility

Presented at Minorities in Engineering Division(MIND) Technical Session 5

The digital accessibility of handwritten STEM materials remains a barrier to inclusive
engineering education, particularly for students who rely on screen readers and text-to-speech technologies. The goal of this study is to better understand the challenges involved in converting handwritten mathematical and other STEM content into structured digital formats. We present a comprehensive evaluation of the current state of Handwritten Mathematical Expression Recognition (HMER) models, with a focused analysis of the MinerU-2.5 Vision Language Model’s performance. By analyzing the equations from multimodal engineering and computing lecture slides, we identify a critical performance gap between symbolic recognition and structural representation. Our findings show that while the model achieves a high rate (87.0%) in recovering mathematical meaning, spatial failures such as collapsed formatting, misinterpreted
sub/superscripts, and dropped summation bounds result in a lower overall interpretability rate (72.4%). We formalize these challenges into a three-part taxonomy of obstacles to symbolic accuracy: structural and spatial failures in preserving 2D arrangement or delimiters, environmental interference from multimodal elements like arrows or overlapping boundaries, and semantic inaccuracies as a result of ambiguous character substitutions. Furthermore, our layout analysis reveals a 66.5% accuracy rate for page-level decomposition, with a 9.6% failure rate primarily caused by multimodal complexity. These obstacles are also formalized into three categories based on whether they stem from issues with type-casting, environmental boundaries, or hierarchical structure. The paper concludes that modern HMER systems are currently bottlenecked more by spatial hierarchy modeling than raw symbol detection. The classification of obstacles in converting handwritten mathematical and other STEM content in engineering courses
to LaTeX helps increase the digital accessibility of STEM resources by providing a roadmap for developing practical guidelines for instruction design, benchmarking, and tool development.

Authors

Advita Gelli University of Illinois at Urbana - Champaign
Alan Tao University of Illinois at Urbana - Champaign
YANGYANG ZHANG University of Illinois at Urbana - Champaign
Nikitha Sundaram
Sonika Tamilarasan The University of Illinois at Chicago
Jonathan Hogg University of Illinois at Urbana - Champaign
Yang Victoria Shao University of Illinois at Urbana - Champaign [biography]

Yang V. Shao is a Teaching Associate Professor in electrical and computer engineering department at University of Illinois Urbana-Champaign (UIUC). She earned her Ph.D. degrees in electrical engineering from Chinese Academy of Sciences, China. She has worked with University of New Mexico before joining UIUC where she developed some graduate courses on Electromagnetics. Dr. Shao has research interests in curriculum development, assessment, student retention and student success in engineering, developing innovative ways of merging engineering fundamentals and research applications.
Dr. Chrysafis Vogiatzis http://orcid.org/0000-0003-0787-9380 University of Illinois at Urbana - Champaign [biography]

Dr. Chrysafis Vogiatzis is a teaching associate professor for the Department of Industrial and Enterprise Systems Engineering at the University of Illinois Urbana-Champaign. Prior to that, Dr. Vogiatzis was an assistant professor at North Carolina Agricultural and Technical State University. His current research interests lie in network optimization and combinatorial optimization, along with their vast applications in modern socio-technical and biological systems. He is serving as the faculty advisor of the Institute of Industrial and Systems Engineers, and was awarded the 2019, 2023, and 2024 Faculty Advisor award for the North-Central region of IISE. Dr. Vogiatzis was awarded ASEE IL/IN Teacher of the Year in 2023.
Dr. Pablo D Robles-Granda University of Illinois at Urbana - Champaign [biography]

Pablo Robles-Granda is a Teaching Assistant Professor at the University of Illinois at Urbana-Champaign.
Prof. Lawrence Angrave http://orcid.org/0000-0001-9762-7181 University of Illinois at Urbana - Champaign [biography]

Dr. Lawrence Angrave is an award-winning computer science Teaching Professor at the University of Illinois Urbana-Champaign. He creates and researches new opportunities for accessible and inclusive equitable education.
Dr. Hongye Liu University of Illinois at Urbana - Champaign [biography]

Hongye Liu is a Teaching Assistant Professor in the Dept. of Computer Science in UIUC. She is interested in education research to help students with disability and broaden participation in computer science.

Note

The full paper will be available to logged in and registered conference attendees once the conference starts on June 21, 2026, and to all visitors after the conference ends on June 24, 2026

« View session