Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Rusakov, Eugen, Sudholt, Sebastian, Wolf, Fabian, Fink, Gernot A.
Format: Preprint
Veröffentlicht: 2018
Schlagworte:
Online-Zugang:https://arxiv.org/abs/1806.10866
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913261209780224
author Rusakov, Eugen
Sudholt, Sebastian
Wolf, Fabian
Fink, Gernot A.
author_facet Rusakov, Eugen
Sudholt, Sebastian
Wolf, Fabian
Fink, Gernot A.
contents The goal in word spotting is to retrieve parts of document images which are relevant with respect to a certain user-defined query. The recent past has seen attribute-based Convolutional Neural Networks take over this field of research. As is common for other fields of computer vision, the CNNs used for this task are already considerably deep. The question that arises, however, is: How complex does a CNN have to be for word spotting? Are increasingly deeper models giving increasingly better results or does performance behave asymptotically for these architectures? On the other hand, can similar results be obtained with a much smaller CNN? The goal of this paper is to give an answer to these questions. Therefore, the recently successful TPP-PHOCNet will be compared to a Residual Network, a Densely Connected Convolutional Network and a LeNet architecture empirically. As will be seen in the evaluation, a complex model can be beneficial for word spotting on harder tasks such as the IAM Offline Database but gives no advantage for easier benchmarks such as the George Washington Database.
format Preprint
id arxiv_https___arxiv_org_abs_1806_10866
institution arXiv
publishDate 2018
record_format arxiv
spellingShingle Exploring Architectures for CNN-Based Word Spotting
Rusakov, Eugen
Sudholt, Sebastian
Wolf, Fabian
Fink, Gernot A.
Computer Vision and Pattern Recognition
The goal in word spotting is to retrieve parts of document images which are relevant with respect to a certain user-defined query. The recent past has seen attribute-based Convolutional Neural Networks take over this field of research. As is common for other fields of computer vision, the CNNs used for this task are already considerably deep. The question that arises, however, is: How complex does a CNN have to be for word spotting? Are increasingly deeper models giving increasingly better results or does performance behave asymptotically for these architectures? On the other hand, can similar results be obtained with a much smaller CNN? The goal of this paper is to give an answer to these questions. Therefore, the recently successful TPP-PHOCNet will be compared to a Residual Network, a Densely Connected Convolutional Network and a LeNet architecture empirically. As will be seen in the evaluation, a complex model can be beneficial for word spotting on harder tasks such as the IAM Offline Database but gives no advantage for easier benchmarks such as the George Washington Database.
title Exploring Architectures for CNN-Based Word Spotting
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/1806.10866