AutoMeta tool for reference

[juhoinkinen]

Found this AutoMeta tool for metadata extraction:

AutoMeta is a metadata extractor tool for scanned Electronic Theses and Dissertations (ETDs). It has been built to extract seven metadata fields from the cover page of scanned ETDs. These fields are: Title, Author, Advisor, University, Degree, Program, and Year. It utilize learning based model such as CRF model with text-based and visual-based features.

We could check the quality of its results compare to Meteor and LLM based methods.

[juhoinkinen]

The group behind AutoMeta has published some papers about theses and dissertations.

Here is a presentation about ETDSuite/ETDMiner (a library including AutoMeta), which aims to segment and parse theses and dissertations. The segmentation employs a multimodal model to classify pages into 13 categories.

[juhoinkinen]

The poster paper "A Heuristic Baseline Method for Metadata Extraction from Scanned Electronic Theses and Dissertations" is most relevant for just metadata extraction, although from 2020:

The process started with converting scanned pages into images and then
text files by applying OCR tools. Then a series of carefully designed
regular expressions for each field is applied, capturing patterns
for seven metadata fields: titles, authors, years, degrees, academic
programs, institutions, and advisors. The method is evaluated on a
ground truth dataset comprised of rectified metadata provided by
the Virginia Tech and MIT libraries.

[osma]

This more recent paper by (mostly) the same authors goes into more detail:
MetaEnhance: Metadata Quality Improvement for Electronic Theses and Dissertations of University Libraries
https://ieeexplore.ieee.org/document/10265916
preprint https://arxiv.org/abs/2303.17661