Please use this identifier to cite or link to this item:
Title: Visual search and applications using convolutional neural networks
Authors: Tanojo, Hosiana Elvirya
Keywords: DRNTU::Engineering
Issue Date: 2018
Abstract: Convolutional Neural Networks (CNN) is one of the most prominent deep learning architecture in performing large scale visual intelligent tasks. CNN can be used for many purposes based on the data given. One of the purpose is to recognise tattoo images to assist in better law enforcement process. Tattoo has been used to identify criminal suspects because it is a biometric characteristic that makes it easier to narrow down and identify victims. This project focuses on developing image recognition application by using CNN on tattoo images dataset, particularly on tattoo classification and detection. Classification will identify if there is any tattoo exists in an image and detection will locate the identified tattoo. A total of 2000 images were used to train the network by using CNN and Fast R- CNN code for classification and detection respectively. Tattoo classification achieved accuracy of 93.33% and tattoo detection achieved mean Average Precision (mAP) or Intersection over Union (IoU) value of 70%. Both results are promising and achieved generally higher accuracy than similar research or standard. Future research can be done to enhance both the classification and detection feature.
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
FYP Final Report- Hosiana.pdf
  Restricted Access
14.38 MBAdobe PDFView/Open

Page view(s) 50

checked on Oct 21, 2020

Download(s) 50

checked on Oct 21, 2020

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.