Mutaz Elradi S. Saeed
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Mutaz Elradi S. Saeed has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
Ali Tahir
College of Engineering and Computer Science, Jazan University
2 shared publicationsBin Gao
University of Electronic Science and Technology of China
2 shared publicationsFagen Li
University of Electronic Science and Technology of China
2 shared publicationsGuiYun Tian
University of Electronic Science and Technology of China
2 shared publications- Guiyun Tian
Chongqing University of Technology
2 shared publications - Bin Gao
School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
1 shared publication
Latest Publications
- Open Access
- 1 October 2026
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 21 (10) , e0358764
Standardization of body weight distribution using a Force-Sensitive Resistor (FSR) matrix and foot analytics
Balanced body weight distribution across both feet is essential for maintaining musculoskeletal health and minimizing postural and ambulatory issues. Conventionally foot weight distribution is measured using specialized clinical equipment which are expensive and limits routine screening. Therefore, in this work, we present a standardized foot weight distribution model having a custom 32 × 32 force-sensing resistor (FSR) matrix based embedded system. A total of 4,000 volunteers participated in the Physical Fitness Assessment (PFA). Of these, 2,200 who passed the musculoskeletal fitness criteria were selected for foot pressure data collection. From these data, we derived a standard weight distribution pattern across foot regions: heel (45–55%), midfoot (10–15%), metatarsals (17–27%), and toes (8–13%). These standardized weight distribution patterns are used to identify deviations associated with musculoskeletal conditions that support early personalized rehabilitation and training strategies. The proposed model is implemented on a real-time embedded system and cloud platform. The embedded system uses a K230 RISC-V processor and performs real-time data acquisition, on-board data processing, and visualization of foot weight distribution. The application running on cloud platform performs statistical analysis and applies a classical support vector machine (SVM) for classifying abnormal pressure patterns. A Convolution Neural Network (CNN) based model is developed that learns fit and un-fit foot patterns from the 32 × 32 pressure maps. The model results show improved detection of dysfunctional weight distribution. The CNN approach shows that AI model can perform advanced predictive analytics in gait and posture assessment. The analysis is limited to static standing posture and does not include dynamic gait assessment. The normative ranges are derived primarily from subjects up to approximately 80–90 kg, as no musculoskeletally fit participants were identified in the heavier weight groups (90–120 kg); therefore, generalization to heavier individuals requires caution and further data collection.
- 1 October 2026
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 21 (10) , e0358764
Open AccessStandardization of body weight distribution using a Force-Sensitive Resistor (FSR) matrix and foot analytics
Balanced body weight distribution across both feet is essential for maintaining musculoskeletal health and minimizing postural and ambulatory issues. Conventionally foot weight distribution is measured using specialized clinical equipment which are expensive and limits routine screening. Therefore, in this work, we present a standardized foot weight distribution model having a custom 32 × 32 force-sensing resistor (FSR) matrix based embedded system. A total of 4,000 volunteers participated in the Physical Fitness Assessment (PFA). Of these, 2,200 who passed the musculoskeletal fitness criteria were selected for foot pressure data collection. From these data, we derived a standard weight distribution pattern across foot regions: heel (45–55%), midfoot (10–15%), metatarsals (17–27%), and toes (8–13%). These standardized weight distribution patterns are used to identify deviations associated with musculoskeletal conditions that support early personalized rehabilitation and training strategies. The proposed model is implemented on a real-time embedded system and cloud platform. The embedded system uses a K230 RISC-V processor and performs real-time data acquisition, on-board data processing, and visualization of foot weight distribution. The application running on cloud platform performs statistical analysis and applies a classical support vector machine (SVM) for classifying abnormal pressure patterns. A Convolution Neural Network (CNN) based model is developed that learns fit and un-fit foot patterns from the 32 × 32 pressure maps. The model results show improved detection of dysfunctional weight distribution. The CNN approach shows that AI model can perform advanced predictive analytics in gait and posture assessment. The analysis is limited to static standing posture and does not include dynamic gait assessment. The normative ranges are derived primarily from subjects up to approximately 80–90 kg, as no musculoskeletally fit participants were identified in the heavier weight groups (90–120 kg); therefore, generalization to heavier individuals requires caution and further data collection.
- Open Access
- 4 November 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (11) , e0335341
Attention to detail: A conditional multi-head transformer for traffic sign recognition
The challenge of traffic sign detection and recognition for driving vehicles has become more critical with recent advances in autonomous and assisted driving technologies. Although object recognition problems, particularly traffic sign recognition, have been extensively studied, most Vision Transformer (ViT) models still rely on static attention mechanisms with fixed projection matrices (Q, K, and V). Using this mechanism limits the ViTs to handle real-world problems such as object detection and traffic sign recognition, etc. Problems, such as partially or fully obscured signs, changes in illumination, and weather conditions, result in subpar feature extraction, which compounds the misclassification problem. To overcome this challenge, a Conditional Visual Transformer (CViT) is proposed in this research, which dynamically adapts feature aggregation, Q, K, and V projections, as well as attention-based mechanisms, based on the input sign type. Its main component consists of a controlled failure deep learning model using a CViT that targets specific types of traffic signs through varying feature extraction and attention adjustments, resulting in high classification performance and minimizing misclassifications. Furthermore, an adaptive gating technique is employed that optimally adjusts the projection matrix across different traffic signs. The proposed CViT achieved an overall accuracy of 99.87%, with a Micro Precision of 99.07%, a Macro Recall of 94.3%, and a Macro F1 Score of 99.07%, respectively. These results demonstrate the potential of CViT to improve both the efficiency and reliability of traffic sign recognition in autonomous driving applications.
- 4 November 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (11) , e0335341
Open AccessAttention to detail: A conditional multi-head transformer for traffic sign recognition
The challenge of traffic sign detection and recognition for driving vehicles has become more critical with recent advances in autonomous and assisted driving technologies. Although object recognition problems, particularly traffic sign recognition, have been extensively studied, most Vision Transformer (ViT) models still rely on static attention mechanisms with fixed projection matrices (Q, K, and V). Using this mechanism limits the ViTs to handle real-world problems such as object detection and traffic sign recognition, etc. Problems, such as partially or fully obscured signs, changes in illumination, and weather conditions, result in subpar feature extraction, which compounds the misclassification problem. To overcome this challenge, a Conditional Visual Transformer (CViT) is proposed in this research, which dynamically adapts feature aggregation, Q, K, and V projections, as well as attention-based mechanisms, based on the input sign type. Its main component consists of a controlled failure deep learning model using a CViT that targets specific types of traffic signs through varying feature extraction and attention adjustments, resulting in high classification performance and minimizing misclassifications. Furthermore, an adaptive gating technique is employed that optimally adjusts the projection matrix across different traffic signs. The proposed CViT achieved an overall accuracy of 99.87%, with a Micro Precision of 99.07%, a Macro Recall of 94.3%, and a Macro F1 Score of 99.07%, respectively. These results demonstrate the potential of CViT to improve both the efficiency and reliability of traffic sign recognition in autonomous driving applications.
- 18 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 5 (6) , 4926-4944
Remote Authentication Schemes for Wireless Body Area Networks Based on the Internet of Things
Internet of Things is a new technology which offers enormous applications that make people’s lives more convenient and enhances cities’ development. In particular, smart healthcare applications in IoTs have been receiving increasing attention for industrial and academic research. However, due to the sensitiveness of medical information, security and privacy issues in IoTs healthcare systems are very important. Designing an efficient secure scheme with less computation time and energy consumption is a critical challenge in IoTs healthcare systems. In this paper, a lightweight online/offline certificateless signature (L-OOCLS) is proposed, then a heterogeneous remote anonymous authentication protocol (HRAAP) is designed to enable remote WBANs users to anonymously enjoy healthcare service based on the Internet of Things applications. The proposed L-OOCLS scheme is proven secure in random oracle model and the proposed HRAAP can resist various types of attacks. Compared with the existing relevant schemes, the proposed HRAAP achieves less computation overhead as well as less power consumption on WBANs client. In addition, to nicely meet the application in the IoTs, an application scenario is given.
- 18 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 5 (6) , 4926-4944
Remote Authentication Schemes for Wireless Body Area Networks Based on the Internet of Things
Internet of Things is a new technology which offers enormous applications that make people’s lives more convenient and enhances cities’ development. In particular, smart healthcare applications in IoTs have been receiving increasing attention for industrial and academic research. However, due to the sensitiveness of medical information, security and privacy issues in IoTs healthcare systems are very important. Designing an efficient secure scheme with less computation time and energy consumption is a critical challenge in IoTs healthcare systems. In this paper, a lightweight online/offline certificateless signature (L-OOCLS) is proposed, then a heterogeneous remote anonymous authentication protocol (HRAAP) is designed to enable remote WBANs users to anonymously enjoy healthcare service based on the Internet of Things applications. The proposed L-OOCLS scheme is proven secure in random oracle model and the proposed HRAAP can resist various types of attacks. Compared with the existing relevant schemes, the proposed HRAAP achieves less computation overhead as well as less power consumption on WBANs client. In addition, to nicely meet the application in the IoTs, an application scenario is given.
- 10 March 2018
- Published by Springer Nature in Wireless Networks
- Vol. 25 (6) , 3081-3101
AKAIoTs: authenticated key agreement for Internet of Things
- 10 March 2018
- Published by Springer Nature in Wireless Networks
- Vol. 25 (6) , 3081-3101
AKAIoTs: authenticated key agreement for Internet of Things
- 9 May 2017
- Published by Springer Nature in Wireless Networks
- Vol. 24 (8) , 3141-3160
HOOSC: heterogeneous online/offline signcryption for the Internet of Things
- 9 May 2017
- Published by Springer Nature in Wireless Networks
- Vol. 24 (8) , 3141-3160
HOOSC: heterogeneous online/offline signcryption for the Internet of Things
- Top Cited PapersOpen Access
- 1 May 2016
- Published by Elsevier in Composite Structures
- Vol. 143, 352-361
Low energy impact damage detection in CFRP using eddy current pulsed thermography
- 1 May 2016
- Published by Elsevier in Composite Structures
- Vol. 143, 352-361
Top Cited PapersOpen AccessLow energy impact damage detection in CFRP using eddy current pulsed thermography