Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy
Counting and detecting occluded faces in a crowd is a challenging task in computer vision. In this paper, we propose a new approach to face detection-based crowd estimation under significant occlusion and head posture variations. Most state-of-the-art face detectors cannot detect excessively occlude...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Sciendo
2023-03-01
|
Series: | International Journal of Applied Mathematics and Computer Science |
Subjects: | |
Online Access: | https://doi.org/10.34768/amcs-2023-0001 |
_version_ | 1797848533216264192 |
---|---|
author | Kian Ara Rouhollah Matiolanski Andrzej Grega Michał Dziech Andrzej Baran Remigiusz |
author_facet | Kian Ara Rouhollah Matiolanski Andrzej Grega Michał Dziech Andrzej Baran Remigiusz |
author_sort | Kian Ara Rouhollah |
collection | DOAJ |
description | Counting and detecting occluded faces in a crowd is a challenging task in computer vision. In this paper, we propose a new approach to face detection-based crowd estimation under significant occlusion and head posture variations. Most state-of-the-art face detectors cannot detect excessively occluded faces. To address the problem, an improved approach to training various detectors is described. To obtain a reasonable evaluation of our solution, we trained and tested the model on our substantially occluded data set. The dataset contains images with up to 90 degrees out-of-plane rotation and faces with 25%, 50%, and 75% occlusion levels. In this study, we trained the proposed model on 48,000 images obtained from our dataset consisting of 19 crowd scenes. To evaluate the model, we used 109 images with face counts ranging from 21 to 905 and with an average of 145 individuals per image. Detecting faces in crowded scenes with the underlying challenges cannot be addressed using a single face detection method. Therefore, a robust method for counting visible faces in a crowd is proposed by combining different traditional machine learning and convolutional neural network algorithms. Utilizing a network based on the VGGNet architecture, the proposed algorithm outperforms various state-of-the-art algorithms in detecting faces ‘in-the-wild’. In addition, the performance of the proposed approach is evaluated on publicly available datasets containing in-plane/out-of-plane rotation images as well as images with various lighting changes. The proposed approach achieved similar or higher accuracy. |
first_indexed | 2024-04-09T18:29:05Z |
format | Article |
id | doaj.art-cf2dfbf0c21b4a85a814947f59320039 |
institution | Directory Open Access Journal |
issn | 2083-8492 |
language | English |
last_indexed | 2024-04-09T18:29:05Z |
publishDate | 2023-03-01 |
publisher | Sciendo |
record_format | Article |
series | International Journal of Applied Mathematics and Computer Science |
spelling | doaj.art-cf2dfbf0c21b4a85a814947f593200392023-04-11T17:28:19ZengSciendoInternational Journal of Applied Mathematics and Computer Science2083-84922023-03-0133172010.34768/amcs-2023-0001Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window StrategyKian Ara Rouhollah0Matiolanski Andrzej1Grega Michał2Dziech Andrzej3Baran Remigiusz41Institute of Telecommunications, AGH University of Science and Technology al. Mickiewicza 30, 30-059Kraków, Poland1Institute of Telecommunications, AGH University of Science and Technology al. Mickiewicza 30, 30-059Kraków, Poland1Institute of Telecommunications, AGH University of Science and Technology al. Mickiewicza 30, 30-059Kraków, Poland1Institute of Telecommunications, AGH University of Science and Technology al. Mickiewicza 30, 30-059Kraków, Poland2Department of Computer Science, Electronics and Electrical Engineering Kielce University of Technology ul. Żeromskiego 5, 25-369Kielce, PolandCounting and detecting occluded faces in a crowd is a challenging task in computer vision. In this paper, we propose a new approach to face detection-based crowd estimation under significant occlusion and head posture variations. Most state-of-the-art face detectors cannot detect excessively occluded faces. To address the problem, an improved approach to training various detectors is described. To obtain a reasonable evaluation of our solution, we trained and tested the model on our substantially occluded data set. The dataset contains images with up to 90 degrees out-of-plane rotation and faces with 25%, 50%, and 75% occlusion levels. In this study, we trained the proposed model on 48,000 images obtained from our dataset consisting of 19 crowd scenes. To evaluate the model, we used 109 images with face counts ranging from 21 to 905 and with an average of 145 individuals per image. Detecting faces in crowded scenes with the underlying challenges cannot be addressed using a single face detection method. Therefore, a robust method for counting visible faces in a crowd is proposed by combining different traditional machine learning and convolutional neural network algorithms. Utilizing a network based on the VGGNet architecture, the proposed algorithm outperforms various state-of-the-art algorithms in detecting faces ‘in-the-wild’. In addition, the performance of the proposed approach is evaluated on publicly available datasets containing in-plane/out-of-plane rotation images as well as images with various lighting changes. The proposed approach achieved similar or higher accuracy.https://doi.org/10.34768/amcs-2023-0001crowd densityface detectionhead pose variationsvarious lighting conditionsocclusion |
spellingShingle | Kian Ara Rouhollah Matiolanski Andrzej Grega Michał Dziech Andrzej Baran Remigiusz Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy International Journal of Applied Mathematics and Computer Science crowd density face detection head pose variations various lighting conditions occlusion |
title | Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy |
title_full | Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy |
title_fullStr | Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy |
title_full_unstemmed | Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy |
title_short | Efficient Face Detection Based Crowd Density Estimation using Convolutional Neural Networks and an Improved Sliding Window Strategy |
title_sort | efficient face detection based crowd density estimation using convolutional neural networks and an improved sliding window strategy |
topic | crowd density face detection head pose variations various lighting conditions occlusion |
url | https://doi.org/10.34768/amcs-2023-0001 |
work_keys_str_mv | AT kianararouhollah efficientfacedetectionbasedcrowddensityestimationusingconvolutionalneuralnetworksandanimprovedslidingwindowstrategy AT matiolanskiandrzej efficientfacedetectionbasedcrowddensityestimationusingconvolutionalneuralnetworksandanimprovedslidingwindowstrategy AT gregamichał efficientfacedetectionbasedcrowddensityestimationusingconvolutionalneuralnetworksandanimprovedslidingwindowstrategy AT dziechandrzej efficientfacedetectionbasedcrowddensityestimationusingconvolutionalneuralnetworksandanimprovedslidingwindowstrategy AT baranremigiusz efficientfacedetectionbasedcrowddensityestimationusingconvolutionalneuralnetworksandanimprovedslidingwindowstrategy |