Historical Document Analysis 1 -- BoundaryNet: An Attentive Deep Network with Fast Marching Distance Maps for Semi-automatic Layout Annotation -- Pho(SC)Net: An Approach Towards Zero-shot Word Image Recognition in Historical Documents -- Versailles-FP dataset: Wall Detection in Ancient Floor Plans -- Graph Convolutional Neural Networks for Learning Attribute Representations for Word Spotting -- Context Aware Generation of Cuneiform Signs -- Adaptive Scaling for Archival Table Structure Recognition -- Document Analysis Systems -- LGPMA: Complicated Table Structure Recognition with Local and Global Pyramid Mask Alignment -- VSR: A Unified Framework for Document Layout Analysis combining Vision, Semantics and Relations -- Layout-Parser: A Unified Toolkit for Deep Learning Based Document Image Analysis -- Understanding and Mitigating the Impact of Model Compression for Document Image Classification -- Hierarchical and Multimodal Classification of Images from Soil Remediation Reports -- Competition and Collaboration in Document Analysis and Recognition -- Handwriting Recognition -- 2D Self-Attention Convolutional Recurrent Network for Offline Handwritten Text Recognition -- Handwritten Text Recognition with Convolutional Prototype Network and Most Aligned Frame Based CTC Training -- Online Spatio-Temporal 3D Convolutional Neural Network for Early Recognition of Handwritten Gestures -- Mix-Up Augmentation for Oracle Character Recognition with Imbalanced Data Distribution -- Radical Composition Network for Chinese Character Generation -- SmartPatch: Improving Handwritten Word Imitation with Patch Discriminators -- Scene Text Detection and Recognition -- Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition -- Text Detection by Jointly Learning Character and Word Regions -- Vision Transformer for Fast and Efficient Scene Text Recognition -- Look, Read and Ask: Learning to Ask Questions by Reading Text in Images -- CATNet: Scene Text Recognition Guided by Concatenating Augmented Text Features -- Explore Hierarchical Relations Reasoning and Global Information Aggregation -- Historical Document Analysis 2 -- One-Model Ensemble-Learning for Text Recognition of Historical Printings -- On the use of attention in deep learning based denoising method for ancient Cham inscription images -- Visual FUDGE: Form Understanding via Dynamic Graph Editing -- Annotation-Free Character Detection in Historical Vietnamese Stele Images -- Document Image Processing -- DocReader: Bounding-Box Free Training of a Document Information Extraction Model -- Document Dewarping with Control Points -- Unknown-box Approximation to Improve Optical Character Recognition Performance -- Document Domain Randomization for Deep Learning Document Layout Extraction -- NLP for Document Understanding -- Distilling the Documents for Relation Extraction by Topic Segmentation -- LAMBERT: Layout-Aware Language Modeling for Information Extraction -- ViBERTgrid: A Jointly Trained Multi-Modal 2D Document Representation for Key Information Extraction from Documents -- Kleister: Key Information Extraction Datasets Involving Long Documents with Complex Layouts -- Graphics, Diagram, and Math Recognition -- Towards an efficient framework for Data Extraction from Chart Images -- Geometric Object 3D Reconstruction From Single Line Drawings Image Based on a Network for Classification and Sketch Extraction -- DiagramNet: Hand-drawn Diagram Recognition using Visual Arrow-relation Detection -- Formula Citation Graph Based Mathematical Information Retrieval.

This four-volume set of LNCS 12821, LNCS 12822, LNCS 12823 and LNCS 12824, constitutes the refereed proceedings of the 16th International Conference on Document Analysis and Recognition, ICDAR 2021, held in Lausanne, Switzerland in September 2021. The 182 full papers were carefully reviewed and selected from 340 submissions, and are presented with 13 competition reports. The papers are organized into the following topical sections: historical document analysis, document analysis systems, handwriting recognition, scene text detection and recognition, document image processing, natural language processing (NLP) for document understanding, and graphics, diagram and math recognition.