Works · Projects and code
From manuscript to machine, across the world's scripts.
The projects Mem carried out or took part in during 2025 and 2026, alongside foundations, universities, libraries and research infrastructures. For each one: what we did, the numbers, the stack and where the code lives.
The code of the ITSERR work packages is public on GitHub under Mem-dh. Our other repositories stay private.
All works
-
Corpus Juris Canonici
Systematic correction of the dataset against the reference critical edition (E. Friedberg, Leipzig 1879–1881) across the Decretum Gratiani and the Liber Extra, and a proof of concept for automatic layout segmentation of pre-modern printed pages.
- 172annotated pages of the Liber Extra
- YOLOv8m-segfine-tuned layout model
- YOLOv8
- eScriptorium
- ALTO XML
-
Image Tagger
A web application for drawing and classifying bounding boxes on page images, linking each OCR segment from Google Vision AI to its area of the page and saving everything back to the source spreadsheets.
- 1,900+Arabic page images verified
- Flask + canvaslightweight, runs locally
- Flask
- JavaScript
- openpyxl
- Google Vision AI
-
Plorabunt
An automated pipeline that collects, transcribes and annotates videos about violent events at places of worship: multilingual transcription with Whisper and metadata augmentation with Mistral served through vLLM.
- 227source rows with video
- 187unique URLs processed
- yt-dlp
- Whisper
- vLLM
- Mistral
-
Sanctuaria
OCR digitisation of hagiographic legends on the foundation of sanctuaries, from printed Italian volumes of the 16th to the early 18th century, with manual revision and metadata for the Sanctuaria tool.
- 40legends digitised
- 6regions, from Calabria to Tuscany
- Transkribus
- Google Gemini
-
Papal bulls
Digitisation of papal bulls and a synoptic comparison of their transcriptions across three reference editions.
- 1,162pages digitised
- 3editions compared: MGH, CCO, EFR
- OCR
- Synoptic comparison
-
Annotation platform and video meta-scraping
Two proofs of concept: a platform for annotating and building high-quality datasets of texts in non-Latin scripts, for cataloguing and for historical-religious research; and a controlled meta-scraping system that collects and structures some 250 videos on religious subjects.
- 2proofs of concept
- ~250religious videos structured and annotated
- Annotation
- Dataset building
- Meta-scraping
Service project · no public code -
Datasets for ancient languages
Linguistic services and annotated datasets for ancient languages, including Chinese and Latin, delivered with CoCp (CITI) in Reggio Emilia for the ITSERR research infrastructure.
- Chinese · Latinancient languages covered
- Annotated datasets
- Linguistic services
Service project · no public code -
EPRIM · Essentials and Pitfalls in Research Infrastructure Management
A feasibility study and training materials on the management of research infrastructures, with particular reference to ESFRI: governance, financial sustainability, operations and legal frameworks, plus an operational alignment report template (ROAR). Delivered in two workshops with Michele Carraglia (ISTI-CNR).
- 2training workshops
- ROARoperational alignment template
- Research infrastructures
- ESFRI
- Training
Service project · no public code -
Deliverable support for ITSERR
Specialist support to Webion for the technical documentation of ITSERR deliverables D1 and D2: collecting the material from the work packages, drafting narrative sections, providing templates and validating the documents before delivery.
- D1 · D2deliverables supported
- WP5–8work packages covered
- Documentation
- Project reporting
Service project · no public code -
Automated cataloguing of the Arabic book
A semi-automated cataloguing system for Arabic-language Islamic collections: OCR extraction, classification with an Islamic-sciences taxonomy and bilingual MARC 21 records, with semantic enrichment of the La Pira Classification and interoperability tests with the National Library Service (SBN).
- 0.92F1 on the validation set
- 34,353Qabas-SAMA lemma pairs mapped
- OCR
- LLM distillation
- MARC 21 XML
- ISO 233-2
-
The photographic archive of Don Milani
AI and computer-vision tools calibrated on the photographic archive of Don Lorenzo Milani for the project Digitalizzare Milani (DIGIμ): automatic image metadata, expert verification, and images and metadata adapted for touch-screen consultation in libraries and museums.
- AI + visionimage metadata pipeline
- Touch screenconsultation in libraries and museums
- Computer vision
- AI metadata
- Expert validation
Service project · no public code -
RESILIENCE Data Ecosystem
Preparation, FAIRification and documentation of consortium-partner datasets for ingestion into ReIReSearch, the Zenodo RESILIENCE Community and the infrastructure's website: source-analysis reports, a FAIR tutorial and the project reports D1–D3, with the LIBIS team (KU Leuven).
- schema.orgJSON-LD metadata mapping
- D1–D3reports co-authored
- schema.org
- JSON-LD
- Zenodo
- MARC
Service project · no public code -
Research, digitisation and editorial services
A continuing framework agreement with the Foundation for Religious Sciences: historical-religious research, high-definition digitisation, classification and digital archiving, software and AI tools for non-Latin scripts, and the editing and layout of journals and book series.
- HDdigitisation of documents and volumes
- Journalsediting, proofreading, layout
- Digitisation
- Editorial
- AI fine-tuning
Service project · no public code -
Editorial and redactional services
From the editing and digitisation of texts, with volumes converted from print into editable files, to the full graphic and editorial production of journals, volumes and series, in print and digital formats.
- Print + digitaljournals, volumes, series
- Editing
- Layout
- Digitisation
Service project · no public code -
Computational component for an ERC proposal
Digital-humanities methodology for an ERC proposal in philosophy: a multilingual corpus with OCR/HTR and TEI encoding, a bilingual SKOS thesaurus, a six-axis annotation framework and embedding analysis under validity controls.
- EN · DEmultilingual corpus
- 6annotation axes
- TEI
- SKOS
- Embeddings
Service project · no public code -
Digitisation and long-term preservation for libraries
Partner of the Universal Civil Service projects Distese di parole and Intrecci di racconti: advice on digitisation and long-term preservation for the G. Dossetti Library in Bologna, the G. La Pira Library in Palermo and the library of the Adria-Rovigo seminary.
- 2civil service projects
- 3libraries supported
- Digitisation
- Long-term preservation
Service project · no public code
No works in this category yet.
Clients and partners
Foundations, universities, libraries, publishers and research infrastructures we work with, or with whom we are preparing new projects.
- FSCIREFoundation for Religious Sciences, Bologna
- UNIMOREUniversity of Modena and Reggio Emilia · ITSERR
- RESILIENCEEuropean research infrastructure for religious studies
- KU Leuven · LIBISData ecosystem of RESILIENCE
- Giorgio La Pira LibraryPalermo · Digital Maktaba
- Il PorticoEDB and Marietti, publishers
- University of FlorenceDepartment of Letters and Philosophy
- ASSRIAssociation for the development of religious sciences in Italy
- WebionITSERR work packages 5–8
- CoCp · CITIReggio Emilia · linguistic services
- ISTI-CNRTraining on research infrastructures
- SCUBOUniversal Civil Service, Bologna