Sefism is free for learners. We are here to help you grow, one step at a time.Join for free
Sefism.
Other

Lecture PDF Accessibility Checker and Accessible HTML Converter for Universities

Build a tool that checks the PDF handouts and slides teachers share on a university learning portal, explains accessibility failures in plain language, and converts the worst ones into accessible HTML with real headings, reading order, alt text prompts and readable tables. Most lecture PDFs are exported from slides with no structure, which locks out blind and low-vision students. The project combines document parsing, a teacher-friendly report and a study with students who use screen readers.

Topics
Difficulty

Intermediate

A small team, or one strong student willing to learn something new.
Effort

1 semester · 2 students

Deliverables

7 to ship

4 optional extras

Suggested stack

PythonPyMuPDF and pdfplumberveraPDF (command line)FastAPIReactNVDA for testing
A suggestion, not a requirement. Swap anything for what you already know.

What you should ship

  • A checker that runs veraPDF and custom checks and reports results in plain language a teacher understands
  • Custom checks for untagged PDFs, scanned pages with no text, missing document language, images without alt text and slide titles that are not headings
  • A converter from slide-style PDFs to accessible HTML with headings, lists and logical reading order
  • An alt text workflow where the teacher writes or approves descriptions for each image, with optional AI drafts clearly marked
  • A batch mode that checks every PDF in a course folder and ranks files by severity
  • A study with at least three screen reader users comparing original PDFs and converted HTML on real tasks
  • An audit of at least 100 real lecture PDFs from your department (with permission) summarised by fault type

If you have time left

  • A plug-in for Moodle or the university's learning portal
  • Equation handling with MathML output
  • Table structure detection for data tables
  • Urdu text extraction and direction handling

The problem: course material that a screen reader cannot follow

Teachers share most course material as PDFs: slides exported from PowerPoint, scanned chapters and handouts typed in Word. To a blind or low-vision student using a screen reader, many of these files are a jumble. Slides exported without tags have no headings, so the student cannot jump between topics. Text boxes are read in the order they were drawn, not the order they make sense. Diagrams have no description. Scanned pages are just images with no text at all.

Universities increasingly accept students with disabilities, but very few teachers know how to make an accessible document, and nobody checks.

What already exists and the gap

  • PAC (PDF Accessibility Checker) by axes4 is a free, widely respected checker for PDF/UA and the Matterhorn Protocol. It is aimed at specialists.
  • veraPDF is an open source validator for PDF/UA and PDF/A. Its reports use the language of the standard, which most teachers cannot act on.
  • Adobe Acrobat has an accessibility checker and remediation tools, but remediation is slow and needs a paid licence.
  • SciA11y from the Allen Institute for AI converts scientific papers into accessible HTML. It is built for research papers, not slide decks and handouts.

The gap is a tool that speaks to teachers in plain language, works on a whole course folder at once and converts lecture-style PDFs into something a screen reader user can actually navigate.

Inspiration

Improving the Accessibility of Scientific Documents (Wang et al., 2021) assessed 11,397 scientific PDFs and found only 2.4 percent met all of the accessibility criteria the authors defined. Their SciA11y system converts papers into accessible HTML with tagged headings, a table of contents and linked references. This project takes the same idea, convert rather than repair, and applies it to the teaching material students use every week.

What you build

  • Teachers upload a PDF or point the tool at a course folder. They get a short report: "This file has no headings, 12 images have no description, page 7 is a scan with no text."
  • The converter turns slide-style PDFs into a single accessible HTML page with one heading per slide, lists preserved and images flagged for descriptions.
  • Teachers add or approve alt text for each image in a simple form. If an AI draft is offered, it is labelled as a draft for review.
  • Students get a clean HTML version they can navigate by headings.

The technically interesting part

  • Recovering structure. Slide PDFs have no tags, but they have patterns: large text near the top is a title, bullets are lists, page footers repeat. Use font size, position and repetition to infer structure, and measure how often you get it right.
  • Reading order. Text boxes in a slide can be in any order in the file. Build a reading-order algorithm using layout (columns, top-to-bottom, left-to-right) and compare with human ordering.
  • Translating standard-speak. Map veraPDF's rule failures to plain-language explanations and fixes a teacher can follow in PowerPoint or Word.
  • Honest testing. Screen reader users judge the result, not the team.

How to evaluate it

  • Structure accuracy: on 50 hand-labelled slide decks, precision and recall of headings and reading order.
  • Report clarity: can teachers fix a document using only your report? Time them and count fixes.
  • User study: screen reader users complete tasks such as "find the definition of normalisation in lecture 4" with the original PDF and the converted HTML; record time and success.
  • Audit: the share of your department's PDFs with each fault type.

Suggested plan

  • Weeks 1 to 3: learn PDF structure and the PDF/UA basics; run PAC and veraPDF on sample files; recruit study participants.
  • Weeks 4 to 7: checker and plain-language report.
  • Weeks 8 to 11: structure recovery and HTML converter.
  • Weeks 12 to 13: alt text workflow, batch mode and the department audit.
  • Weeks 14 to 15: user study, report and demo.

Path to a product

Universities and online course providers face growing accessibility expectations, and converting old material is a large job. A tool sold or donated to university disability offices, or offered as a plug-in for learning portals, is realistic. It is also a portfolio piece that shows document processing, accessibility and user research together.

Scope warning

Do not write a PDF parser or validator; use PyMuPDF, pdfplumber and veraPDF. Do not try to fix PDFs in place; convert to HTML. Scanned pages need OCR, which can be a stretch goal. Get permission before auditing anyone's course material, and report results by fault type, not by teacher.

Learn what you need on Sefism

Get help with this idea

The hard parts are recovering structure from untagged slides and running a respectful study with screen reader users. Tauseef can help you plan both and keep the scope to slides and handouts. If you are new here, create a free account first.

Ideas and guidance, not finished projects

These are project ideas and scoping guidance, published free for students to use as a starting point. I do not build, write, or sell final-year projects, and I do not complete coursework for anyone. Take an idea, make it yours, and build it.

Not sure this one fits you?

I help students pick an idea that matches their skills and their deadline, then scope it down until it is finishable. Guidance only, never done for you.

Discuss this idea
Work with me

Stuck on something specific?

Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.

Book a free sessionSee all services

Follow along

New writing, resources and project ideas land here first.