Depth estimation is an important task in autonomous driving, and usually needs special types of sensors or multiple cameras. In this paper, we propose a novel approach to monocular depth estimation based on two other cheaper annotation tasks: semantic segmentation and prediction of a single vanishing point without the need for ground truth depth data. In a Manhattan-world assumption with a single vanishing point, only one vanishing point exists and represents the end of the scene extension on the z-axis. Depending on semantic segmentation prediction, we set hand-crafted rules to determine the depth of each pixel in the scene depending on its label and its spatial position with regard to the vanishing point. We train two convolutional neural networks (CNNs): a semantic segmentation CNN and a vanishing point prediction CNN. We then fuse the results obtained from the two networks using the hand-crafted rules, which are defined based on single-view geometry rules by taking into consideration the label of the pixel and the nature of the object obtained by the segmentation model. Extensive experiments were done using the KITTI and Cityscapes benchmark datasets. The proposed model achieves impressive performance in semantic segmentation (mean intersection over union of 82.20%) and vanishing point estimation (mean absolute error of 1.87). Monocular depth estimation achieved a relative absolute error of 0.070 with the KITTI dataset and 0.289 with the Cityscapes dataset, outperforming many state-of-the-art methods in depth estimation and semantic segmentation at 10 frames per second.

Due to the wide range of intensive computational applications and ubiquitous connectivity of the Internet of Things (IoT) paradigms, it has become crucial to develop a new platform that can achieve low delay, high network throughput, and enhanced quality of service (QoS). This paper proposes a joint user association, service caching, and task offloading strategy to reduce delay and enhance users’ QoS in multi-tier communication and multi-tier edge computing heterogeneous networks (HetNets). The considered system model consists of multi-users with different tasks and service data sizes communicating in a heterogeneous network of one massive multiple-input multiple-output (M-MIMO) macro base station and some small base stations. The proposed work investigates user association, power allocation, optimum service data caching, and task offloading strategies at the computing network edges. Thereby, the …

Retained surgical items (RSIs) pose significant risks to patients and healthcare professionals, prompting extensive efforts to reduce their incidence. RSIs are objects inadvertently left within patients’ bodies after surgery, which can lead to severe consequences such as infections and death. The repercussions highlight the critical need to address this issue. Machine learning (ML) and deep learning (DL) have displayed considerable potential for enhancing the prevention of RSIs through heightened precision and decreased reliance on human involvement. ML techniques are finding an expanding number of applications in medicine, ranging from automated imaging analysis to diagnosis. DL has enabled substantial advances in the prediction capabilities of computers by combining the availability of massive volumes of data with extremely effective learning algorithms. This paper reviews and evaluates recently

Cancelable template protection techniques are indispensable to provide essential security and privacy privileges in biometric systems. This paper introduces an efficient cancelable palmprint recognition technique based on multi-level transformations. Gabor filtering with feature remapping is introduced for extracting highly discriminative features. Histogram remapping is applied to nonlinearly transform the downsampled Gabor features to be normally distributed. This feature-remapping step is proposed to enhance the discriminatory power and alleviate the effect of feature variability and image artifacts. Comb filtering is applied to the mapped features as a first protection layer. To provide security guarantees against linkability attacks, index-based locality-sensitive hashing (LSH) is introduced as a second protection layer to transform the comb-filtered mapped real-valued features into maximum-ranked indices …

The growing demand for high-quality video requires complex coding techniques that cost resource consumption and increase encoding time which represents a challenge for real-time processing on Embedded Systems. Kvazaar and x265 encoders are two efficient implementations of the High-Efficient Video Coding (HEVC) standard. In this paper, the performance of All Intra Kvazaar and x265 encoders on the Nvidia Jetson platform was evaluated using two coding configurations; highspeed preset and high-quality preset. In our work, we used two scenarios, first, the two encoders were run on the CPU, and based on the average encoding time Kvazaar proved to be 65.44% and 69.4% faster than x265 with 1.88% and 0.6% BD-rate improvement over x265 at high-speed and high-quality preset, respectively. In the second scenario, the two encoders were run on the GPU of the Nvidia Jetson, and the results show the …

Smart cities (SCs) were founded on the basis of the Internet of Vehicles paradigm, aiming to enhance the quality of life. Despite not having a unique composition, the intelligent transportation services (ITS) is a key component in most of the SCs. The ITS integrates the new communication systems along with the traditional transportation systems, forming a universal network of static and vehicular entities communicating over wireless broadcast channels. In addition to its vulnerability to co-channel interference (CCI), eavesdropping, and fading, the vehicle-to-vehicle communication channels could be subjected to shadowed fading. That is due to the probabilistic obstruction by big vehicles. Hence, degrading the signal-to-noise ratio at the receiving vehicle severely. This work evaluates the secrecy performance of a practical cooperative vehicular relaying network in terms of its secrecy outage probability (SOP). Due to …

Unmanned aerial vehicles (UAVs) are expected to be integrated into future wireless networks to offer services, especially in unreachable or congested areas. To improve the spectral efficiency, non-orthogonal multiple access (NOMA) scheme can be utilised within the UAV communication to allow more users to be covered and associated. The performance of the NOMA-UAVs network is governed by several factors including power allocation, user association and pairing methods. This paper presents an approach that uses multi-armed bandit (MAB) and two-sided matching frameworks to maximize the throughput of multi-UAV-assisted NOMA networks in a decentralized manner. The approach enables the UAVs to propose to the ground users (GUs) without explicit cooperation among the UAVs while the GUs can accept or reject the proposals. To this end, we propose a modified Thompson sampling algorithm that we named …

The stochastic routing transportation network problem presents significant challenges due to uncertainty in travel times, real-time variability, and limited sensor data availability. Traditional adaptive routing strategies, which rely on real-time travel time updates, may lead to suboptimal decisions due to dynamic traffic fluctuations. This study introduces a novel routing framework that integrates traffic sensor data augmentation and deep learning techniques to improve the reliability of route selection and network observability. The proposed methodology consists of four components: stochastic traffic assignment, multi-objective route generation, optimal traffic sensor location selection, and deep learning-based traffic flow estimation. The framework employs a traffic sensor location problem formulation to determine the minimum required sensor deployment while ensuring an accurate network-wide traffic estimation. A Stacked Sparse Auto-Encoder (SAE) deep learning model is then used to infer unobserved link flows, enhancing the observability of stochastic traffic conditions. By addressing the gap between limited sensor availability and complete network observability, this study offers a scalable and cost-effective solution for real-time traffic management and vehicle routing optimization. The results confirm that the proposed data-driven approach significantly reduces the need for sensor deployment while maintaining high accuracy in traffic flow predictions.

This research aims to study the impact of Zycobond and Terrasil (liquid nano-acrylic co-polymer) as additives with cement and hydrated lime for stabilizing highway subgrade silty soil. Laboratory tests were conducted to identify the most suitable rehabilitation technique for subgrade silty soils with high plasticity at optimal moisture levels, and to assess the related performance characteristics (i.e., unconfined compressive strength (UCS), California bearing ratio (CBR), and permeability) for the implementation of environmentally friendly road pavement systems. Atterberg's limits, hydrometer analysis, UCS, direct shear, AASHTO, and unified classification systems were used to identify the fundamental characteristics of the reference soil. Three types of soil modifiers were considered: Portland cement (PC), hydrated lime (Ca(OH)2), and nano polymer solution. To determine the best ratio for the nano polymer solution, three different percentages of PC and Ca (OH) 2 were selected by soil weight (1%, 3%, and 5%) at OMC. 1%Ca(OH)2 and 3%PC were more suitable according to the UCS, and the Atterberg limits. The results indicated that the reference soil's maximum compressive strength (qu) improved when treated with either 3%PC or 1%Ca(OH)2 combined with the nanopolymer solution. The compressive strength increased by 67.41 percent and 28.35 percent, respectively. The permeability of the soil modified with 3% PC and 1% Ca(OH)2 using the nanopolymer solution decreased by 87.55 percent and 93.3 percent compared with the reference soil, respectively. It was found that the increase in CBR in 3% PC-modified soil treated with a nano polymer solution was 378.66 percent, whereas in 1% Ca(OH)2-modified soil treated with the same solution, it was 231.17 percent compared to the reference soil.