The automatedexaminer.
Photograph a handwritten answer script. AutoEx reads it — English, Arabic or equations — scores it against the keywords you set, explains every mark, and reports on the whole class.
How a page becomes a mark
The paper's process flow, runningTry two stages, right here
Nothing leaves your browserStage 1 — segmentation, in your browser
Thresholding…
Stages 3–4 — correction and matching
Testing image segmentation with more than 2 words at once
Plurals match by stem; a misspelling within one edit (two for long words) matches fuzzily. Exact hits are solid, fuzzy ones dashed.
What a marked script looks like
Computed live from the sample answers
Missing on this page: palisade, surface area, veins, light intensity, carbon dioxide concentration — the chips say exactly where every mark went.
…and the class, six scripts later
mean 53.3% · s.d. 37.5| Student | Q1 /10 | Q2 /8 | Q3 /7 | Total | % | Grade |
|---|---|---|---|---|---|---|
| Aisha Al-Harbi | 10 | 8 | 7 | 25 | 100 | S |
| Omar Siddiqui | 10 | 4 | 2 | 16 | 64 | C |
| Lin Chen | 6 | 3 | 2 | 11 | 44 | E |
| Faisal Al-Otaibi | 2 | 0 | 0 | 2 | 8 | F |
| Maria Santos | 8 | 7 | 7 | 22 | 88 | A |
| Yusuf Demir | 2 | 2 | 0 | 4 | 16 | F |
Three kinds of paper
All nine samples are in the examiner
Numbered answers, each question marked on its own section.
Right-to-left pages read as written; diacritics, alef forms and the definite article folded before matching.
Formula, substitution and answer each earn marks — x², ½at² and 117.6 N are matched as written.
What the examiner gets
Each question lists the keywords found, as the student wrote them, and how they matched. Change a keyword and the whole class re-scores.
Equal splits a question's marks evenly across its keywords. Custom weights the concepts that matter. Marks never exceed a question's maximum.
Mean, standard deviation, grades (fixed bands or the class's own quartiles), a linear model, and the keywords the class missed most.
A vision model in the cloud for cursive and messy scripts, or on-device Tesseract when a page must not leave the room. Every transcription can be corrected by hand.
Exams, scripts and marks live in your browser's storage. No accounts, no database, no student data on our servers.
One click exports the mark sheet — student, per-question marks, total, percentage, grade — ready for a spreadsheet or the institution's system.
The research behind it
AutoEx began as a university course project and became a peer-reviewed paper: handwriting is segmented with contour analysis, read by a CNN–RNN network trained with CTC loss (cascaded with Apple's Vision framework), corrected with an NLP pipeline and scored by keyword weightage. Its reported word error rate was 5.6%.
This web version keeps the method and swaps the parts that needed a laptop for parts that run anywhere: the segmenter runs in the browser, the reader is a vision model, the R interface became this.
Design of automated model for inspecting and evaluating handwritten answer scripts: A pedagogical approach with NLP and deep learning
Alexandria Engineering Journal, Volume 108 (2024), pp. 764–788. Open access (CC BY-NC-ND).
AutoEx for your institution
The demo here is free to try. For a school, college or examination board we set it up as a system — a row at a time, on your ground — and price it for your size, not by the page.
Mark a class now
Nine sample exams — English, Arabic and equations — with the questions, keywords and scripts already in them. Pick one and press read.
Open the examiner