Dr. Shadrack Mahenge (PhD) | Xidian University (original) (raw)

Uploads

Papers by Dr. Shadrack Mahenge (PhD)

2022 26th International Computer Science and Engineering Conference (ICSEC)

Proceedings of the 8th International Conference on Computing and Artificial Intelligence

Cracks on road surfaces causes inconveniences to drivers and passengers and may cause mechanical ... more Cracks on road surfaces causes inconveniences to drivers and passengers and may cause mechanical failure or even accidents. Good Road condition plays an important role in quick transportation of goods and services from one place to another and acts as a catalyst for the economic development. Road surfaces need to be maintained in good condition to ensure the safety of road users. Road damage detection is important for Structural Health Monitoring (SHM). Traditional manual inspection is normally performed through human visualization which is time consuming, expensive, dangerous because of the passing vehicles, suffers from subjective judgment of the inspector and pose difficulties in keeping records for future road maintenance and repair. The rapid emergency and development of AI has stimulated many experts to automate the process of crack detection through computer vision (CV) technology, though most of these studies faces challenge on getting good detection accuracy. In this study a novel modified U-Net Architecture for image classification and segmentation is proposed to detect cracks on the road surfaces by using detection and classification of the road images to determine whether they represent cracks or not.

Extensive experiments are conducted on three publicly available road crack datasets to evaluate the performance of our proposed model, The performance of the proposed Modified U-Net architecture was verified with respect to different performance metrics such as accuracy, precision, recall and f1 score. Qualitative and Quantitative comparisons experimental results of the proposed approach were also compared with existing state of the art U-Net architectures. It can be inferred from results that the proposed approach achieves superior performance in terms of detection accuracy.

2022 The 6th International Conference on Compute and Data Analysis

Automatic detection of road cracks is one of the significant aspects of road maintenance systems.... more Automatic detection of road cracks is one of the significant aspects of road maintenance systems. However, it involves a lot of complexities to accurately identify the cracks because of various reasons such as heterogeneity of traffic conditions, dynamic environmental conditions and variabilities in multiple parameters. With the advancements in Deep Learning (DL) techniques, this research proposes a DL-based road crack detection model which combines two effective techniques namely RCNN and GAN. A novel RCNN-GAN deep architecture is proposed with reduced layers to improve the detection accuracy. The performance of the proposed approach was verified with respect to different performance metrics such as accuracy, precision, recall, and f1-score. The experimental results of the proposed approach were also compared to various state-of-the art deep learning algorithms. It can be inferred from results that the proposed approach achieves superior performance in terms of detection accuracy and other performance indicators.

Secure Password Authentication Scheme by using Cryptographic Key Exchange in Servers, 2017

In current paper, we present a privateness keeping data-leak detection approach to solve the hind... more In current paper, we present a privateness keeping data-leak detection approach to solve the hindrance where a precise set of touchy knowledge digests is used in detection. The abilities of our method is that it allows the information owner to soundly delegate the detection operation to a semi-sincere provider without revealing the touchy information to the supplier. We describe how ISPs can present their buyers knowledge-leak detection as an add on carrier with strong privacy ensures. In this paper, we recall a situation where two servers cooperate to authenticate a patron and if one server is compromised, the attacker still cannot faux to be the customer with the expertise from the compromised server. Current solutions for 2-server PAKE are either symmetric in the experience that two peer servers equally make a contribution to the authentication or uneven within the sense that one server authenticates the consumer with the help of an extra server. This paper offers a symmetric solution for 2-server PAKE, where the purchaser can set up distinct cryptographic keys with the 2 servers, respectively.

2021 The 5th International Conference on Video and Image Processing, 2021

Computer vision (CV) based inspection has recently attracted considerable attention and is progre... more Computer vision (CV) based inspection has recently attracted considerable attention and is progressively replacing traditional visual inspection which is subject to poor accuracy, high subjectivity, and inefficiency. This paper, benefiting from hybrid structures of multichannel parallel convolutional neural networks (pCNNs), introduces a unique deep learning framework for road crack detection. Ideally, CNN-based frameworks require relatively huge computing resources for accurate image analysis. However, the portability objective of this work necessitates the utilization of low-power processing units. To that purpose, we propose robust deep representation learning for Road Crack Detection (RoCDe) which uses multichannel pCNNs. Bayesian optimization algorithm (BOA) was used to optimize the multichannel pCNNs training with the fewest possible neural network (NN) layers to achieve maximum accuracy, improved efficiency, and minimum processing time. The CV training was done using two distinct optimizers namely Adam and RELU on a sufficiently available dataset through image preprocessing and data augmentation. Experimental results show that the proposed algorithm can achieve high accuracy around 95% in crack detection, which is good enough to replace human inspections normally conducted on-site. This is largely due to well-calibrated predictive uncertainty estimates (WPUE). The effectiveness of the proposed model is demonstrated and validated empirically via extensive experiments and rigorous evaluation on large scale real-world datasets. Furthermore, the performance of hybrid CNNs is compared with state-of-the-art NN models, and the results pro- vides remarkable difference in success level, proving the strength of multichannel pCNNs.

International Journal of Pervasive Computing and Communications, 2021

Purpose In the purpose of the section, the cracks that are in the construction domain may be comm... more Purpose In the purpose of the section, the cracks that are in the construction domain may be common and usually fixed with the human inspection which is at the visible range, but for the cracks which may exist at the distant place for the human eye in the same building but can be captured with the camera. If the crack size is quite big can be visible but few cracks will be present due to the flaws in the construction of walls which needs authentic information and confirmation about it for the successful completion of the wall cracks, as these cracks in the wall will result in the structure collapse. Design/methodology/approach In the modern era of digital image processing, it has captured the importance in all the domain of engineering and all the fields irrespective of the division of the engineering, hence, in this research study an attempt is made to deal with the wall cracks which are found or searched during the building inspection process, in the present context in association...

2022 26th International Computer Science and Engineering Conference (ICSEC)

Proceedings of the 8th International Conference on Computing and Artificial Intelligence

Cracks on road surfaces causes inconveniences to drivers and passengers and may cause mechanical ... more Cracks on road surfaces causes inconveniences to drivers and passengers and may cause mechanical failure or even accidents. Good Road condition plays an important role in quick transportation of goods and services from one place to another and acts as a catalyst for the economic development. Road surfaces need to be maintained in good condition to ensure the safety of road users. Road damage detection is important for Structural Health Monitoring (SHM). Traditional manual inspection is normally performed through human visualization which is time consuming, expensive, dangerous because of the passing vehicles, suffers from subjective judgment of the inspector and pose difficulties in keeping records for future road maintenance and repair. The rapid emergency and development of AI has stimulated many experts to automate the process of crack detection through computer vision (CV) technology, though most of these studies faces challenge on getting good detection accuracy. In this study a novel modified U-Net Architecture for image classification and segmentation is proposed to detect cracks on the road surfaces by using detection and classification of the road images to determine whether they represent cracks or not.

Extensive experiments are conducted on three publicly available road crack datasets to evaluate the performance of our proposed model, The performance of the proposed Modified U-Net architecture was verified with respect to different performance metrics such as accuracy, precision, recall and f1 score. Qualitative and Quantitative comparisons experimental results of the proposed approach were also compared with existing state of the art U-Net architectures. It can be inferred from results that the proposed approach achieves superior performance in terms of detection accuracy.

2022 The 6th International Conference on Compute and Data Analysis

Automatic detection of road cracks is one of the significant aspects of road maintenance systems.... more Automatic detection of road cracks is one of the significant aspects of road maintenance systems. However, it involves a lot of complexities to accurately identify the cracks because of various reasons such as heterogeneity of traffic conditions, dynamic environmental conditions and variabilities in multiple parameters. With the advancements in Deep Learning (DL) techniques, this research proposes a DL-based road crack detection model which combines two effective techniques namely RCNN and GAN. A novel RCNN-GAN deep architecture is proposed with reduced layers to improve the detection accuracy. The performance of the proposed approach was verified with respect to different performance metrics such as accuracy, precision, recall, and f1-score. The experimental results of the proposed approach were also compared to various state-of-the art deep learning algorithms. It can be inferred from results that the proposed approach achieves superior performance in terms of detection accuracy and other performance indicators.

Secure Password Authentication Scheme by using Cryptographic Key Exchange in Servers, 2017

In current paper, we present a privateness keeping data-leak detection approach to solve the hind... more In current paper, we present a privateness keeping data-leak detection approach to solve the hindrance where a precise set of touchy knowledge digests is used in detection. The abilities of our method is that it allows the information owner to soundly delegate the detection operation to a semi-sincere provider without revealing the touchy information to the supplier. We describe how ISPs can present their buyers knowledge-leak detection as an add on carrier with strong privacy ensures. In this paper, we recall a situation where two servers cooperate to authenticate a patron and if one server is compromised, the attacker still cannot faux to be the customer with the expertise from the compromised server. Current solutions for 2-server PAKE are either symmetric in the experience that two peer servers equally make a contribution to the authentication or uneven within the sense that one server authenticates the consumer with the help of an extra server. This paper offers a symmetric solution for 2-server PAKE, where the purchaser can set up distinct cryptographic keys with the 2 servers, respectively.

2021 The 5th International Conference on Video and Image Processing, 2021

Computer vision (CV) based inspection has recently attracted considerable attention and is progre... more Computer vision (CV) based inspection has recently attracted considerable attention and is progressively replacing traditional visual inspection which is subject to poor accuracy, high subjectivity, and inefficiency. This paper, benefiting from hybrid structures of multichannel parallel convolutional neural networks (pCNNs), introduces a unique deep learning framework for road crack detection. Ideally, CNN-based frameworks require relatively huge computing resources for accurate image analysis. However, the portability objective of this work necessitates the utilization of low-power processing units. To that purpose, we propose robust deep representation learning for Road Crack Detection (RoCDe) which uses multichannel pCNNs. Bayesian optimization algorithm (BOA) was used to optimize the multichannel pCNNs training with the fewest possible neural network (NN) layers to achieve maximum accuracy, improved efficiency, and minimum processing time. The CV training was done using two distinct optimizers namely Adam and RELU on a sufficiently available dataset through image preprocessing and data augmentation. Experimental results show that the proposed algorithm can achieve high accuracy around 95% in crack detection, which is good enough to replace human inspections normally conducted on-site. This is largely due to well-calibrated predictive uncertainty estimates (WPUE). The effectiveness of the proposed model is demonstrated and validated empirically via extensive experiments and rigorous evaluation on large scale real-world datasets. Furthermore, the performance of hybrid CNNs is compared with state-of-the-art NN models, and the results pro- vides remarkable difference in success level, proving the strength of multichannel pCNNs.

International Journal of Pervasive Computing and Communications, 2021

Purpose In the purpose of the section, the cracks that are in the construction domain may be comm... more Purpose In the purpose of the section, the cracks that are in the construction domain may be common and usually fixed with the human inspection which is at the visible range, but for the cracks which may exist at the distant place for the human eye in the same building but can be captured with the camera. If the crack size is quite big can be visible but few cracks will be present due to the flaws in the construction of walls which needs authentic information and confirmation about it for the successful completion of the wall cracks, as these cracks in the wall will result in the structure collapse. Design/methodology/approach In the modern era of digital image processing, it has captured the importance in all the domain of engineering and all the fields irrespective of the division of the engineering, hence, in this research study an attempt is made to deal with the wall cracks which are found or searched during the building inspection process, in the present context in association...