Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions

As computer vision and machine learning advance, face detection has become a major focus. Face recognition has several methods and models. Every implementation starts with face detection. Haar Cascades and Multi-task Cascaded Convolutional Networks (MTCNN) are compared for facial pose variation robu...

Full description

Bibliographic Details
Main Authors: Naser, Omer Abdulhaleem, Mumtazah, Sharifah, Samsudin, Khairulmizam, Hanafi, Marsyita, Shafie, Siti Mariam, Zamri, Nor Zarina
Format: Article
Language:English
Published: Croatian Communications and Information Society 2025
Online Access:http://psasir.upm.edu.my/id/eprint/120894/
http://psasir.upm.edu.my/id/eprint/120894/1/120894.pdf
_version_ 1848868239759638528
author Naser, Omer Abdulhaleem
Mumtazah, Sharifah
Samsudin, Khairulmizam
Hanafi, Marsyita
Shafie, Siti Mariam
Zamri, Nor Zarina
author_facet Naser, Omer Abdulhaleem
Mumtazah, Sharifah
Samsudin, Khairulmizam
Hanafi, Marsyita
Shafie, Siti Mariam
Zamri, Nor Zarina
author_sort Naser, Omer Abdulhaleem
building UPM Institutional Repository
collection Online Access
description As computer vision and machine learning advance, face detection has become a major focus. Face recognition has several methods and models. Every implementation starts with face detection. Haar Cascades and Multi-task Cascaded Convolutional Networks (MTCNN) are compared for facial pose variation robustness. This research will examine how well these two models detect faces in yaw postures from-90 to +90 degrees. Many studies have contrasted these two models, but the yaw poses of faces were not addressed due to the scarcity of datasets with systematic degrees of face orientation. Thus, the UPM face dataset, created at the UPM embedded systems lab using developed equipment to produce high-resolution photographs and a systematic range of face orientations from-90 to 90 degrees, was used to evaluate the range of degrees these two models can reach. UPM includes 100 students with different yaw angles and occlusions (masks, glasses, or both). The results reveal that MTCNN is the best for detecting faces with yaw poses only, masks, glasses, and both at all degrees (-90 to +90) with 100%, 99.9%, 96.4%, and 80% accuracy. Instead, Haar cascades were 92.5%, 67.3%, 80.4%, and 76.3% accurate.
first_indexed 2025-11-15T14:49:14Z
format Article
id upm-120894
institution Universiti Putra Malaysia
institution_category Local University
language English
last_indexed 2025-11-15T14:49:14Z
publishDate 2025
publisher Croatian Communications and Information Society
recordtype eprints
repository_type Digital Repository
spelling upm-1208942025-10-15T02:45:51Z http://psasir.upm.edu.my/id/eprint/120894/ Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions Naser, Omer Abdulhaleem Mumtazah, Sharifah Samsudin, Khairulmizam Hanafi, Marsyita Shafie, Siti Mariam Zamri, Nor Zarina As computer vision and machine learning advance, face detection has become a major focus. Face recognition has several methods and models. Every implementation starts with face detection. Haar Cascades and Multi-task Cascaded Convolutional Networks (MTCNN) are compared for facial pose variation robustness. This research will examine how well these two models detect faces in yaw postures from-90 to +90 degrees. Many studies have contrasted these two models, but the yaw poses of faces were not addressed due to the scarcity of datasets with systematic degrees of face orientation. Thus, the UPM face dataset, created at the UPM embedded systems lab using developed equipment to produce high-resolution photographs and a systematic range of face orientations from-90 to 90 degrees, was used to evaluate the range of degrees these two models can reach. UPM includes 100 students with different yaw angles and occlusions (masks, glasses, or both). The results reveal that MTCNN is the best for detecting faces with yaw poses only, masks, glasses, and both at all degrees (-90 to +90) with 100%, 99.9%, 96.4%, and 80% accuracy. Instead, Haar cascades were 92.5%, 67.3%, 80.4%, and 76.3% accurate. Croatian Communications and Information Society 2025 Article PeerReviewed text en cc_by_nc_nd_4 http://psasir.upm.edu.my/id/eprint/120894/1/120894.pdf Naser, Omer Abdulhaleem and Mumtazah, Sharifah and Samsudin, Khairulmizam and Hanafi, Marsyita and Shafie, Siti Mariam and Zamri, Nor Zarina (2025) Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions. Journal of Communications Software and Systems, 21 (1). pp. 109-119. ISSN 1845-6421; eISSN: 1846-6079 https://jcoms.fesb.unist.hr/10.24138/jcomss-2024-0084/ 10.24138/jcomss-2024-0084
spellingShingle Naser, Omer Abdulhaleem
Mumtazah, Sharifah
Samsudin, Khairulmizam
Hanafi, Marsyita
Shafie, Siti Mariam
Zamri, Nor Zarina
Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title_full Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title_fullStr Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title_full_unstemmed Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title_short Comparative Analysis of MTCNN and Haar Cascades for Face Detection in Images with Variation in Yaw Poses and Facial Occlusions
title_sort comparative analysis of mtcnn and haar cascades for face detection in images with variation in yaw poses and facial occlusions
url http://psasir.upm.edu.my/id/eprint/120894/
http://psasir.upm.edu.my/id/eprint/120894/
http://psasir.upm.edu.my/id/eprint/120894/
http://psasir.upm.edu.my/id/eprint/120894/1/120894.pdf