Teacher Key-Mark Codes for Handwritten Examination Answer Sheets: A Derived Dataset
Description
This dataset is a derivative dataset built upon the original dataset "A Dataset of Digitized Student Examination Papers, Answer Keys, and Manual Evaluations for Automated Grading Research" by Dinesh K P (DOI: 10.17632/sf3kvjwknt.1). Purpose: To detect key teacher marks on each question and output a mark code, and to record the teacher's summary totals written on the answer sheet (total of MCQ and total of all). It contains a single manually annotated label file (available in both XLS and TXT formats), which includes: - Per-question teacher key-mark codes for questions 1–35 - Teacher summary totals: total of MCQ and total of all It also contains a format conversion script (XLS to TXT). No image files are included in this dataset.
Files
Steps to reproduce
This dataset is built from manual annotations. To use or verify this dataset: 1. Download the original images from the original dataset (DOI: 10.17632/sf3kvjwknt.1). 2. The image filenames (e.g., Student_1.pdf) correspond to the `student seq` in the label file `labels/teacher_marks_codes.xls`. 3. The label file contains the manually annotated key-mark codes for questions 1–35 and the teacher's summary totals (`total of MCQ` and `total of all`). 4. Refer to `labels/README.txt` for the detailed coding rules, and compare the labels with the images to understand or reproduce the annotation decisions. 5. For programmatic use, load `labels/teacher_marks_codes.txt`, which contains the same content in a space-separated format.