Header menu link for other important links
X
ResPho(SC)Net: A Zero-Shot Learning Framework for Norwegian Handwritten Word Image Recognition
Aniket Gurav, Joakim Jensen, , Sukalpa Chanda
Published in Springer Science and Business Media Deutschland GmbH
2023
Volume: 14062 LNCS
   
Abstract
Recent advances in deep Convolutional Neural Networks (CNNs) have established them as a premier technique for a wide range of classification tasks, including object recognition, object detection, image segmentation, face recognition and medical image analysis. However, a significant drawback of utilizing CNNs is the requirement for a large amount of annotated data, which may not be feasible in the context of historical document analysis. In light of this, we present a novel CNN-based architecture ResPho(SC)Net, to recognize handwritten word images in a zero-shot learning framework. Our method proposes a modified version of the Phosc(Net) architecture with a much lesser number of trainable parameters. Experiments were conducted on word images from two languages (Norwegian and English) and encouraging results were obtained. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
About the journal
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Science and Business Media Deutschland GmbH
ISSN03029743