Salvato in:
Dettagli Bibliografici
Autori principali: Heidenreich, Hunter, Dalvi, Ratish, Mukku, Rohith, Verma, Nikhil, Pičuljan, Neven
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2408.11981
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913476567367680
author Heidenreich, Hunter
Dalvi, Ratish
Mukku, Rohith
Verma, Nikhil
Pičuljan, Neven
author_facet Heidenreich, Hunter
Dalvi, Ratish
Mukku, Rohith
Verma, Nikhil
Pičuljan, Neven
contents Page Stream Segmentation (PSS) is an essential prerequisite for automated document processing at scale. However, research progress has been limited by the absence of realistic public benchmarks. This paper works towards addressing this gap by introducing TABME++, an enhanced benchmark featuring commercial Optical Character Recognition (OCR) annotations. We evaluate the performance of large language models (LLMs) on PSS, focusing on decoder-based models fine-tuned with parameter-efficient methods. Our results show that decoder-based LLMs outperform smaller multimodal encoders. Through a review of existing PSS research and datasets, we identify key challenges and advancements in the field. Our findings highlight the key importance of robust OCR, providing valuable insights for the development of more effective document processing systems.
format Preprint
id arxiv_https___arxiv_org_abs_2408_11981
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Large Language Models for Page Stream Segmentation
Heidenreich, Hunter
Dalvi, Ratish
Mukku, Rohith
Verma, Nikhil
Pičuljan, Neven
Computation and Language
Page Stream Segmentation (PSS) is an essential prerequisite for automated document processing at scale. However, research progress has been limited by the absence of realistic public benchmarks. This paper works towards addressing this gap by introducing TABME++, an enhanced benchmark featuring commercial Optical Character Recognition (OCR) annotations. We evaluate the performance of large language models (LLMs) on PSS, focusing on decoder-based models fine-tuned with parameter-efficient methods. Our results show that decoder-based LLMs outperform smaller multimodal encoders. Through a review of existing PSS research and datasets, we identify key challenges and advancements in the field. Our findings highlight the key importance of robust OCR, providing valuable insights for the development of more effective document processing systems.
title Large Language Models for Page Stream Segmentation
topic Computation and Language
url https://arxiv.org/abs/2408.11981