文章基本信息

标题：A Survey of Multi-panel Image Segmentation Methods
本地全文：下载
作者：Farrukh Jabeen ; Mushtaq Ali ; Amanullah Baloch 等
期刊名称：Journal of Computers
印刷版ISSN：1796-203X
出版年度：2018
卷号：13
期号：12
页码：1395-1402
DOI：10.17706/jcp.13.12.1395-1402
语种：English
出版社：Academy Publisher
摘要：Multi-panel images play a vital role in describing complicated situations; however, their availability degrades the retrieval accuracy of image retrieval systems. For improving the accuracy of image retrieval systems, the researchers proposed different approaches for the segmentation of multi-panel images into their constituent sub-images. In this paper, five multi-panel image segmentation approaches are reviewed: (1) Manual Segmentation, (2) Inter-panel Boundary Detection, (3) Cluster Based Approach, (4) Sub-panel Label Detection, and (5) Panel Boundary Detection. We discuss the performance metrics used for the evaluation of these proposed approaches. An overview of the datasets used in the literature is also provided. The practical aspects of sub-image separation are finally considered.
关键词：Image segmentation; boundary detection; subpanel label detection; OCR.