Laypa: A Novel Framework for Applying Segmentation Networks to Historical Documents

Stefan Klut; Rutger Koert, van; R.G.H. Sluijter

Laypa: A Novel Framework for Applying Segmentation Networks to Historical Documents

KNAW HUC

Research output: Contribution to conference › Paper › Scientific › peer-review

3 Citations (Scopus)

Abstract

We present novel software to process scans of historical documents to extract their layout information. We do this using a ResNet backbone with a feature pyramid head. We extract region information directly into PageXML. For baseline extraction, we use a two stage processing approach. The software has been applied successfully to several projects. The results show the feasibility to automatically label text lines and regions in historical documents.

Original language	English
Pages	67-72
Number of pages	6
Publication status	Published - 25 Aug 2023
Event	HIP '23: Proceedings of the 7th International Workshop on Historical Document Imaging and Processing - San José, California, San José, United States Duration: 25 Aug 2023 → 26 Aug 2023 https://dl.acm.org/doi/proceedings/10.1145/3604951

Conference

Conference	HIP '23
Abbreviated title	HIP '23
Country/Territory	United States
City	San José
Period	25/08/2023 → 26/08/2023
Internet address	https://dl.acm.org/doi/proceedings/10.1145/3604951

Keywords

Datasets
historical documents
Neural networks
layout analysis

Cite this

@conference{ed2ca673ddf74ce1b38ed90eabbd241b,

title = "Laypa: A Novel Framework for Applying Segmentation Networks to Historical Documents",

abstract = "We present novel software to process scans of historical documents to extract their layout information. We do this using a ResNet backbone with a feature pyramid head. We extract region information directly into PageXML. For baseline extraction, we use a two stage processing approach. The software has been applied successfully to several projects. The results show the feasibility to automatically label text lines and regions in historical documents.",

keywords = "Datasets, historical documents, Neural networks, layout analysis",

author = "Stefan Klut and \{Koert, van\}, Rutger and R.G.H. Sluijter",

year = "2023",

month = aug,

day = "25",

language = "English",

pages = "67--72",

note = "HIP '23 : Proceedings of the 7th International Workshop on Historical Document Imaging and Processing, HIP '23 ; Conference date: 25-08-2023 Through 26-08-2023",

url = "https://dl.acm.org/doi/proceedings/10.1145/3604951",

}

TY - CONF

T1 - Laypa: A Novel Framework for Applying Segmentation Networks to Historical Documents

AU - Klut, Stefan

AU - Koert, van, Rutger

AU - Sluijter, R.G.H.

PY - 2023/8/25

Y1 - 2023/8/25

N2 - We present novel software to process scans of historical documents to extract their layout information. We do this using a ResNet backbone with a feature pyramid head. We extract region information directly into PageXML. For baseline extraction, we use a two stage processing approach. The software has been applied successfully to several projects. The results show the feasibility to automatically label text lines and regions in historical documents.

AB - We present novel software to process scans of historical documents to extract their layout information. We do this using a ResNet backbone with a feature pyramid head. We extract region information directly into PageXML. For baseline extraction, we use a two stage processing approach. The software has been applied successfully to several projects. The results show the feasibility to automatically label text lines and regions in historical documents.

KW - Datasets

KW - historical documents

KW - Neural networks

KW - layout analysis

UR - https://dl.acm.org/doi/abs/10.1145/3604951.3605520

M3 - Paper

SP - 67

EP - 72

T2 - HIP '23

Y2 - 25 August 2023 through 26 August 2023

ER -

Laypa: A Novel Framework for Applying Segmentation Networks to Historical Documents

Abstract

Conference

Keywords

Other files and links

Fingerprint

Cite this