.png){kind=icon}
.png){kind=icon}
Return
Enhancing Algorithm Comprehension for Visually Impaired Individuals With LLM-Based Programming Code Segmentation
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2025.3638804.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The concept of inclusive education increasingly focuses on providing appropriate support for diverse learners, including visually impaired individuals (VIs) who encounter notable difficulties in programming education. Unlike sighted learners who can visually grasp code structure at a glance, VIs depend on screen readers that present code linearly, hindering their ability to comprehend program flow and increasing cognitive load. This study proposes a method that leverages Large Language Model (LLM) to automatically segment the source code into semantically coherent sections along with their corresponding explanatory descriptions. These descriptions are embedded in a screen-reader-compatible block-based programming interface, designed to enhance non-visual code comprehension by providing structured, high-level guidance. An experimental study was conducted with 20 sighted participants blindfolded to simulate non-visual interaction. Participants completed programming tasks using two systems: one with LLM-generated segmentation and one without. The results demonstrate that the segmented system significantly improved task success rates and reduced completion times by over 50% in the most complex task. Additionally, subjective evaluations indicated enhanced usability and comprehension. These findings highlight the potential of LLM-based code segmentation to improve programming accessibility for VIs. Future work includes testing on complex codebases, adding contextual aids, enhancing segmentation flexibility, and conducting broader longitudinal evaluations.
Keywords:
Assistive technologies
code segmentation
cognitive load
human–computer interaction
inclusive education
large language models
non-visual interfaces
programming education
program comprehension
software accessibility
visual impairments
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.6
Papers:
9.8W
Citations:
29.4W