Improved CNN System for Face Mask Recognition DOI Creative Commons
Ammar Hussein Jassim, Ahmed Al-Taie, Amal Sufiuh Ajrash

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: April 19, 2024

Abstract Deep learning, especially convolutional neural networks, has significantly improved performance in computer vision. Therefore, we designed and developed a modified deep network framework for detecting mask facial images sizable synthesized un-synthesized face dataset. The suggested method can be utilized to detect masks any image with low-resolution, different alignments, complex, noisy background by tuning the hyperparameters accurately identify existence of without generating overfitting. experimentally obtained results demonstrate that model exhibits significant efficiency level, achieving 97.39% accuracy, 97.34% precision, 97.41% recall, 97.37% F1-score, 97.4% AUC. empirical have been documented after 35 iterations using optimized hyperparameter settings, those predictive models were trained on 64,398 98% accuracy rate 0.05 loss, proving proposed work's reliability robustness.

Language: Английский

Navigating the face recognition: unleashing the power of few-shot learning through metric-based insights DOI
Sushant Jain, Amit Pundir, Sanjeev Singh

et al.

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: 83(33), P. 79939 - 79961

Published: March 1, 2024

Language: Английский

Citations

1
YOLOv5-S2C2: An Improved Method of Mask Detection Based on Lightweight DOI Creative Commons
Zongyuan Xie, Hongyan Ma, Wei He

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 53306 - 53319

Published: Jan. 1, 2024

Masks play a crucial role in preventing respiratory diseases and have diverse applications national public health industrial safety. Efficient mask-wearing detection systems are essential for ensuring accuracy real-time performance. To overcome the challenges of extensive model calculations, parameter volume, complex hardware deployment current system, lightweight mask with improved YOLOv5 is proposed. Firstly, this study proposes new network—EMA-FasterNet, as backbone network YOLOv5, which reduces computation while preserving information each channel. Secondly, using Depthwise Separable Convolutions (DepthSepConv) to replace some C3 modules Neck further compresses parameters, computation. Finally, prevent missing objects due removal overlapping candidate boxes, use Soft Non-Maximum Suppression (NMS) NMS. Compared YOLOv5s, proposed YOLOv5-S2C2 has high mAP both Dataset I II, parameters reduced by about 56% 57.0%, respectively. The volume 44% original inference speed on GPU 30%. These improvements show that achieves design excellent It also well suited efficient limited computational resources.

Language: Английский

Citations

0
Improved CNN System for Face Mask Recognition DOI Creative Commons
Ammar Hussein Jassim, Ahmed Al-Taie, Amal Sufiuh Ajrash

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: April 19, 2024

Abstract Deep learning, especially convolutional neural networks, has significantly improved performance in computer vision. Therefore, we designed and developed a modified deep network framework for detecting mask facial images sizable synthesized un-synthesized face dataset. The suggested method can be utilized to detect masks any image with low-resolution, different alignments, complex, noisy background by tuning the hyperparameters accurately identify existence of without generating overfitting. experimentally obtained results demonstrate that model exhibits significant efficiency level, achieving 97.39% accuracy, 97.34% precision, 97.41% recall, 97.37% F1-score, 97.4% AUC. empirical have been documented after 35 iterations using optimized hyperparameter settings, those predictive models were trained on 64,398 98% accuracy rate 0.05 loss, proving proposed work's reliability robustness.

Language: Английский

Citations

0