Please use this identifier to cite or link to this item: http://www.aiktcdspace.org:8080/jspui/handle/123456789/2715
Title: Image auto tagging
Authors: Khan, Tabrez
Gouri, Javed M. (15DCO48)
Shaikh, Mohd. Kalam (15DCO71)
Shaikh, Shahbaz A. (15DCO74)
Khan, Mohd. Bilal (15DCO51)
Keywords: Project Report - CE
Issue Date: May-2018
Publisher: AIKTC
Series/Report no.: Accession # PE0439;
Abstract: Due to the advancement in the field of multimedia technologies, there is an increase in the computerized and digital images. An image may contain a tree, house, mountain, etc due to which a real life object can be categorized into multiple categories. There have been several studies on automatic image annotation where they utilize machine learning techniques to an- notate digital images due to its need. Face detection and recognition is already being used in many real world applications. The traditional methods of retrieving an image such as annotating images manually is time-consuming and expensive, especially for an continuously increasing image database. The problem in the existing applications is that it does not tag the other ob- jects present in the pictures, and sometimes they also have a problem with tagging people. In this paper, we propose a system of automatic image annotation using convolutional neural net- works that takes into account the texual queries or keywords and searches for the related in the database. Image auto-tagging is a classification task that aims to tag or label an image by identifying the objects present within the same image. This new system also has an advantage of automatically determine the image on the basis of the keyword entered by the user. It can also be used to improve information content for the description of the image.
Description: In Partial Fulfillment of the Requirement for the Award of Computer engineering
URI: http://www.aiktcdspace.org:8080/jspui/handle/123456789/2715
Appears in Collections:Project Reports - CO

Files in This Item:
File Description SizeFormat 
PE0439.pdf5.57 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.