Please use this identifier to cite or link to this item: https://gnanaganga.inflibnet.ac.in:8443/jspui/handle/123456789/16757
Title: Best Image Processing for Higher Face Detection Rate Using Haar Cascades
Authors: Gupta, Suneet
Singh, Uday
Rawat, Manoj
Kurian, Asha
Mandal, Lopa
Gupta, Praveen
Sardar, Tanvir Habib
Keywords: Detection Rate
Face Detection
Haar Cascades
Haar Features
New Tps
Test-Images
Issue Date: 2024
Publisher: EAI/Springer Innovations in Communication and Computing
Springer Science and Business Media Deutschland GmbH
Citation: pp. 113-128
Abstract: This chapter finds the best image processing technique that when implemented on a group photo as a preprocessing increases the number of faces detected using the Haar feature-based face detection algorithm. Group photo is a photograph or a picture or an image where we have several faces. Many times not all the faces are detected when a face detection algorithm is implemented on any group photo. So, to detect more number of faces, the image is preprocessed with different preprocessing techniques and each resultant image is examined. Haar feature-based face detection is implemented on each of the resultant images and checked whether any new face(s) has/have been detected or not compared to the faces detected with just the gray image (the one without preprocessing). In fact, the same preprocessing technique is implemented on a set of 15 images and the number of new faces detected in each image is noted. Finally, the total number of new faces detected on the entire set of images is obtained. The same process is repeated with other preprocessing techniques. The preprocessing technique that gives the highest number of new faces is selected as the best image processing technique that a group photo should undergo prior to the implementation of the Haar feature face detection algorithm. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
URI: https://doi.org/10.1007/978-3-031-64495-5_9
https://gnanaganga.inflibnet.ac.in:8443/jspui/handle/123456789/16757
ISBN: 9783031644948
ISSN: 2522-8595
Appears in Collections:Conference Papers

Files in This Item:
There are no files associated with this item.


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