Comparison of Haar-like, HOG and LBP approaches for face detection in video sequences

Adouani, Amal; Henia, Wiem Mimoun Ben; Lachiri, Zied

doi:10.1109/ssd.2019.8893214

Cited by 47 publications

(21 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In each single frame, the feature vector is then constructed using this gradient feature vector. Finally, the HOG feature vector is produced by combining all gradient feature vectors derived from different images, and then inputted to the SVM to extract an array of bounding boxes for the human face [17]. Input image and feature vectors generated from the HOG method can be seen in Figure 7.…”

Section: Face Detectionmentioning

confidence: 99%

Cloud-based architecture for face identification with deep learning using convolutional neural network

Herlambang

Buana

Piarsa

2021

IJEECS

View full text Add to dashboard Cite

The use of a face as a biometric to identify a person in order to keep the system safe from an unauthorized person has advantages over other biometric characteristics. The face as a biometric has more structure and a wider area than other biometrics, while can be retrieved in a non-invasive manner. We proposed a cloud-based architecture for face identification with deep learning using convolutional neural network. Face identification in this study used a cloud-based engine with four stages, namely face detection with histogram of oriented gradients (HOG), image enhancement, feature extraction using convolutional neural network, and classification using k-nearest neighbor (KNN), SVM, as well as random forest algorithm. This study conducted a classification experiment with cloud-based architecture using three different datasets, namely Faces94, Faces96 and University of Manchester Institute of Science and Technology (UMIST) face dataset. The results from this study are with the proposed cloud-based architecture, the best accuracy is obtained by KNN algorithm with an accuracy of 99% on Faces94 dataset, 99% accuracy on Faces96 dataset, 97% on UMIST face dataset, and performance of the three algorithms decreased in UMIST face dataset with facial variations from various angles from left to right profile.

show abstract

Section: Face Detectionmentioning

confidence: 99%

Cloud-based architecture for face identification with deep learning using convolutional neural network

Herlambang

Buana

Piarsa

2021

IJEECS

View full text Add to dashboard Cite

show abstract

“…Thus, there are many researches have concentrated this problem. Face recognition system consists some main steps: face detection( [6], [26], [12]), face representation( [7], [8], [9]) and recognition( [22], [23], [24], [25]). Firstly, face detection aims to detect face region in full images.…”

Section: Introductionmentioning

confidence: 99%

“…This work is necessary in order to choose the best suitable face detection model. Secondly, face representation could be divided into two groups: hand craft-based methods( [8], [9]) and deep learning-based methods( [14], [16], [18]). In the first trend, special points on face are characteristic, fixed and distinguish between humans.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Realtime Face Recognition for Intelligent Door System

2020

IJETER

View full text Add to dashboard Cite

This paper focus on in-deep analysis of the majority of both hand-craft based and CNN-based that proposed in states-of-the-art for face detection and face recognition method. This issue has been researched for recent decades as it enjoys many practical applications such as entertainments, security, tracking, human-machine interaction and so on. However, this problem has been faced to many challenges such as complex background, illumination, face pose, states of face, fake face. Which lead to high time cost, and low accuracy system. In recent times, thanks to the cutting-edge technique that give impressive performance on both hand craft-based and deep learning-based method. There have been improvement in detecting face exploiting feature vectors. Therefore, in this research, we comparing face detection model such as Haar Cascade[6], Dlib [19] and YOLO [27]. Then, we consider and compare affects of methods when extract feature vectors. Besides, we try to examine on various classifications to obtain the suitable system. Evaluation results confirm that the best accuracy rate achieves at 97.83%. The proposed method suggests a feasible solution addressing technical issues in using hand craft-based face recognition. Moreover, we deploy a realtime intelligent door system that utilize face recognition results to automatically control door.

show abstract

“…The standard Dlib detector uses histograms of oriented gradients (HOG) features with linear SVMs. Some studies have found that HOG + SVM performs better than Haar cascades or Linear Binary Patterns (LBPs) for face detection in video sequences [10,11]. The multi-task cascaded convolutional neural network (MTCNN) model for face detection [12] consists of three convolutional networks (R-Net, P-Net, and O-Net) and is able to outperform other methods on face detection benchmarks while retaining real-time performance.…”

mentioning

confidence: 99%

Two-Stage Facial Mask Detection Model for Indoor Environments

Zereen

Corraya

Dailey

et al. 2020

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

The COVID-19 pandemic has led to many organizations around the world enforcing face mask rules for personal protection. Manual checking whether individuals entering an organization's premises are wearing masks is cumbersome and possibly confrontational. There has been relatively little work on automatic face mask rule violations thus far. We propose a system for automatic monitoring for face mask rule violations for enterprises. Our method is an efficient two-stage facial mask detection model. The first stage is based on facial landmark extraction and clustering, and the second stage analyzes the clustered nose region. A thorough accuracy evaluation on five types of sample face images (no mask, beard and mustache, single-color mask, multi-color mask, and skin-color mask) finds that the overall accuracy of the two-stage model is an excellent 97.13%, outperforming simpler single-stage models. IntroductionGlobally, at the time of writing, over 36 million people are infected and over 1,056,186 deaths have been caused by the COVID-19 pandemic [1]. The World Health Organization has recommended that mask-wearing can reduce the chances of being infected or spreading COVID-19 by respiratory droplets, which constitute the main vector

show abstract

Comparison of Haar-like, HOG and LBP approaches for face detection in video sequences

Cited by 47 publications

References 30 publications

Cloud-based architecture for face identification with deep learning using convolutional neural network

Cloud-based architecture for face identification with deep learning using convolutional neural network

Realtime Face Recognition for Intelligent Door System

Two-Stage Facial Mask Detection Model for Indoor Environments

Contact Info

Product

Resources

About