Conference Programme


You can find more information on the floorplan and venue here

Day 1: Thursday, September 03, 2026

Oral Session 1 (9:00 am - 10:30 am) Tables, Forms & Structured Extraction - (Session Chair: Jihad El-Sana) Venue: EI7 (Vienna, Austria)
9:00 am - 9:15 am Opening Ceremony (Faisal Shafait)
9:15 am - 10:00 am Keynote 1: JORDY VAN LANDEGHEM - Parse, Reflect, Retrieve, Compile: An Agent Stack for Enterprise Document AI
10:00 am - 10:15 am O1: [Paper 46] POTATR: A Lightweight Image-to-Graph Model for Page-Level Table Extraction
10:15 am - 10:30 am O2: [Paper 42] Structured Handwritten Form Synthesis via Unsupervised Template Recovery and Field-Level Segmentation
10:30 am - 11:00 am Morning Coffee Break
Oral Session 2 (11:00 am - 12:30 pm) Document-Centric AI, Retrieval & Reasoning - (Session Chair: Antoine Doucet) Venue: EI7 (Vienna, Austria)
11:00 am - 11:45 am Keynote 2: NIBAL NAYEF - From Retrieval to Reasoning: The Evolution of RAG-based Systems for Document-Centric AI
11:45 am - 12:00 pm O3: [Paper 36] Modeling "Musical Words" for Efficient Sequence Search in Handwritten Music Manuscripts
12:00 pm - 12:15 pm O4: [Paper 10] Graph-based Aggregation for Writer Identification and Retrieval
12:15 pm - 12:30 pm O5: [Paper 49] Reading Order Inference for Complex Document Layouts
12:30 pm - 1:30 pm Lunch Break
Poster Session I (1:30 pm - 3:00 pm) Venue: Hallway EI7 (Vienna, Austria)
P1: [Paper 12] DocScribe: Handwriting-Aware Document Templating with Compact OCR-Free Models
P2: [Paper 16] Criticality-Aware Evaluation of Table Understanding Systems on Synthetic ESG Documents
P3: [Paper 17] Towards Zero-Shot Domain Generalization for ID Cards Presentation Attack Detection
P4: [Paper 25] PeopleGator: A Multimodal Benchmark for Person Identification in Historical Documents
P5: [Paper 27] Text Pixels Are Not Enough: A Controlled Study of Canvas Sensitivity in VLM-Based OCR
P6: [Paper 29] Scale-Aware Feature-Level Encoding for Historical Document Stamp Detection
P7: [Paper 35] TrOCR for Medieval HTR: A Systematic Ablation Study with Cross-Dataset Validation
P8: [Paper 37] Handwritten Devanagari Character Recognition: Conjunct Consonants Dataset Creation and Benchmarking with Transfer Learning
P9: [Paper 47] A Feasibility Study of Concept-Based XAI Classification Methods on Chart Images
P10: [Paper 51] BZK-LE: A Dataset for Zero-Shot and Few-Shot Archival Document Layout Classification
P11: [Paper 53] Lect-Former: Localized Self-Attention Skip Connections for Whiteboard Content Extraction
P12: [Paper 52] Evaluating Multimodal Large Language Models for End-to-End Diagram Recognition
3:00 pm - 3:30 pm Afternoon Coffee Break
Demo Session (3:30 pm - 4:15 pm) Venue: Hallway EI7 (Vienna, Austria)
D1: CrimeNER Demo: Named-Entity Recognition in the Crime Domain
D2: AIriskEval-edu Demo: Explainable Auditing of Pedagogical Risks in Educational Explanations
D3: EpiSAM Demo: An Interactive System for Character Segmentation and Text-Line Extraction in Challenging Stone Inscriptions
D4: Multi Script OCR System for Handwritten Indic Manuscripts
D5: Litmus: Zero Label Metric Design for AI Pipelines
D6: Reconstructing Historical Manuscripts through MSI: The Potential of Contrast in Assessing Image Quality and Legibility
4:15 pm - 5:00 pm Discussion Group - (Session Chair: Momina Moetesum)

Day 2: Friday, September 04, 2026

Oral Session 3 (9:00 am - 10:30 am) OCR, Layout & Recognition Systems - (Session Chair: Joseph Chazalon) Venue: EI7 (Vienna, Austria)
9:00 am - 9:45 am Keynote 3: BRANDON SMOCK - Why Haven’t We Solved OCR?
9:45 am - 10:00 am O6: [Paper 43] Nutri-TabNet: Detection of Non-planar Nutrition Information Panels in Real-World Food Packaging
10:00 am - 10:15 am O7: [Paper 11] CollabScore OMR: A Hybrid System for Music Score Recognition
10:15 am - 10:30 am O8: [Paper 31] IndicPage-OCR: Robust Low-Resource Adaptation for Multi-Script Indic Handwritten Page Recognition
10:30 am - 11:00 am Morning Coffee Break
Poster Session II (11:00 am - 12:30 pm) Venue: Hallway EI7 (Vienna, Austria)
P13: [Paper 5] Rethinking Text Color Estimation in Screen Documents: A Reliability-Aware Approach
P14: [Paper 20] Decoupled Prototype Memory Mechanism for Complex Table Recognition
P15: [Paper 21] A Recognition-Aware Multimodal Framework for Actual and Intended Text Transcription of Child Handwriting
P16: [Paper 24] Text-Based Classification of Historical Notarial Documents from Noisy HTR Transcripts
P17: [Paper 50] Mind the Gaps: A Curated Benchmark for Form Field Detection
P18: [Paper 14] SciXplain: Structure-Aware Representation Learning for Scientific Method Diagrams
P19: [Paper 9] Decoding Romanian Law: A Fine-Tuned LLM Approach

SP1: YOTO - You Only Train Once: Template-Specific Table Extraction from a Single Example
SP2: French PDF-to-Markdown Benchmark
SP3: From Pen Strokes to Sleep States: Detecting Low-Recovery Days Using Sigma-Lognormal Handwriting Features
SP4: Reconstructing Historical Manuscripts through MSI: The Potential of Contrast in Assessing Image Quality and Legibility
12:30 pm - 1:30 pm Lunch Break
Oral Session 4 (1:30 pm - 3:00 pm) Enterprise and Industrial Document AI - (Session Chair: Nibal Nayef) Venue: EI7 (Vienna, Austria)
1:30 pm - 2:15 pm Keynote 4: Darko Obradovic - Beyond Accuracy: Confidence and Reliability in Industrial DAS
2:15 pm - 2:30 pm O9: [Paper 30] Edge-Aware Relation Prediction for Symbol-Level Structure Refinement in Online Handwritten Mathematical Expressions
2:30 pm - 2:45 pm O10: [Paper 33] End-to-End Camera-Based Analysis of Egyptian Hieroglyphic Textual Artifacts
2:45 pm - 3:00 pm Closing Ceremony and Awards (Adrian Ulges)
3:00 pm - 3:30 pm Afternoon Coffee Break