Paul Rad

Cloud of Things in Smart Agriculture: Intelligent Irrigation Monitoring by Thermal Imaging

Thermal imaging has shown potential in assisting many aspects of smart irrigation management. This article examines key technical and legal issues and requirements supporting the use of Cloud of Things for managing water source-related data prior to discussing potential solutions.

Low-latency software defined network for high performance clouds

Multi-tenant clouds with resource virtualization offer elasticity of resources and elimination of initial cluster setup cost and time for applications. However, poor network performance, performance variation and noisy neighbors are some of the challenges for execution of high performance applications on public clouds. Utilizing these virtualized resources for scientific applications, which have complex communication patterns, require low latency communication mechanisms and rich set of communication constructs. To minimize the virtualization overhead, a novel approach for low latency network for HPC Clouds is proposed and implemented over a multi-technology software defined network. The efficiency of the proposed low-latency Software Defined Networking is analyzed and evaluated for high performance applications. The results of the experiments show that the latest Mellanox FDR InfiniBand interconnect and Mellanox OpenStack plugin gives the best performance for implementing VM-based high performance clouds with large message sizes.

Cloud-based realtime robotic Visual SLAM

Prior work has shown that Visual SLAM (VSLAM) algorithms can successfully be used for realtime processing on local robots. As the data processing requirements increase, due to image size or robot velocity constraints, local processing may no longer be practical. Offloading the VSLAM processing to systems running in a cloud deployment of Robot Operating System (ROS) is proposed as a method for managing increasing processing constraints. The traditional bottleneck with VSLAM performing feature identification and matching across a large database. In this paper, we present a system and algorithms to reduce computational time and storage requirements for feature identification and matching components of VSLAM by offloading the processing to a cloud comprised of a cluster of compute nodes. We compare this new approach to our prior approach where only the local resources of the robot were used, and examine the increase in throughput made possible with this new processing architecture.

ZeroVM: secure distributed processing for big data analytics

A key challenge for any large-scale computation today, whether in “big data” or in handling large-scale web services, has to do with the management of data. In the big data context, the arbitrary separation of storage and computation increases latency and decreases performance. ZeroVM is a lightweight container-based virtualization platform that provides deterministic process execution and isolation. The philosophy behind ZeroVM is to virtualize applications then move the application to the data. This provides the ability to transform or process data in situ, rather than moving data to where the application is located. With the ability to move and execute application next to data, ZeroVM changes the conventional wisdom on infrastructure centric commuting models and enables even more data centric computing models to be used for Big-Data Analytics. The ZeroVM distributed processing framework proposed in this paper presents new opportunities for processing, storing and using data, particularly in big data analytics.

A Next-Generation Secure Cloud-Based Deep Learning License Plate Recognition for Smart Cities

License Plate Recognition System (LPRS) plays a vital role in smart city initiatives such as traffic control, smart parking, toll management and security. In this article, a cloud-based LPRS is addressed in the context of efficiency where accuracy and speed of processing plays a critical role towards its success. Signature-based features technique as a deep convolutional neural network in a cloud platform is proposed for plate localization, character detection and segmentation. Extracting significant features makes the LPRS to adequately recognize the license plate in a challenging situation such as i) congested traffic with multiple plates in the image ii) plate orientation towards brightness, iii) extra information on the plate, iv) distortion due to wear and tear and v) distortion about captured images in bad weather like as hazy images. Furthermore, the deep learning algorithm computed using bare-metal cloud servers with kernels optimized for NVIDIA GPUs, which speed up the training phase of the CNN LPDS algorithm. The experiments and results show the superiority of the performance in both recall and precision and accuracy in comparison with traditional LP detecting systems.

A novel image encryption method to reduce decryption execution time in cloud

Parallel computing on cloud infrastructure had gained tremendous attention and popularity in recent years. In this paper, we propose a new technique to encrypt an image for secure image transmission and parallel decryption using cloud resources. The encoding of the image is done using the FFT conjugate complex property and private random phase key value shared between sender and receiver. First, using FFT complex conjugate property, we construct a two dimensional matrix consisting of real and imaginary portion of the image. Second, we encrypt the complex conjugate matrix using private random phase key value shared between sender and receiver. Detailed simulations have been carried out to test the encryption algorithm. Furthermore, we use cloud resources for the development of an efficient parallel image encryption and decryption algorithm. Significant speedup was achieved using different image size and number of cloud servers.

Deep Learning Control for Complex and Large Scale Cloud Systems

AbstractDeep learning attempts to model high level perceptions in data using deep graph representations and creating models to learn these representations from large-scale unlabeled signals. Efficient unsupervised feature learning is extracted by deep learning algorithms and with multiple processing layers, composed of multiple linear and non-linear transformations. Actual systems become more and more complex with huge numbers of state variables and control of such large and complex systems with chaotic behavior, which needs more information about systems. Deep learning control by discovering continoiusly almost all possible information seems to be a reasonable approach to model and control largescale and complex systems. Recent advancements in machine learning algorithms and platforms are leading to deep learning controllers in real-time applications. The goal of this paper is to describe the concept of deep learning control and explain how cloud fog computing and edge analytics could handle massive amou...

Container based parallelization for faster and reliable image segmentation

In this paper, we describe a scalable and economical architecture for performing container based parallelization to obtain the best possible quantized image using different quantization techniques on the cloud. This approach using containers can be scaled to be used with huge datasets. The quantization techniques used in this paper are fuzzy entropy and genetic algorithm based techniques. Different types of membership functions are used in each technique to calculate the fuzzy entropy. The best possible quantized image is determined using the Structural Similarity Index (SSIM). This is a futuristic approach for solving lengthy repetitive serial problems in a parallel and economical way. As expected the results significantly better than the serial approach.

Image segmentation by multi-level thresholding using genetic algorithm with fuzzy entropy cost functions

In this paper, we describe three new soft computing methods for segmentation of both gray level and color images by using a fuzzy entropy based cost function for the genetic algorithm. The presented methods allow us to find optimized set of parameters for a predefined cost function. Particularly, we found the optimum set of membership functions by maximizing the fuzzy entropy and based on the membership functions. Experimental results of thresholding and comparison of SSIM of thresholded images with different techniques are presented. Results show that the offered method can reliably segment and give better thresholds then Otsu Multi-Level thresholding.

A Privacy-Aware Architecture at the Edge for Autonomous Real-Time Identity Reidentification in Crowds

The capability to perform identity reidentification in a crowd (e.g., from video feeds from a network of cameras, and social media platforms, such as Facebook and Instagram) efficiently and effectively is increasingly important, as evident in recent real-world events (e.g., terrorist attacks on places of mass gatherings in different countries). However, real-time reidentification in a network of cameras, such as those deployed in a smart city, and from other sources, such as social media platforms, remains a challenging task. In this paper, a new embedding algorithm pipeline is presented to extract and administrate the crowd-sourced facial image features (e.g., social media platforms and multicameras in a dense crowd, such as a stadium or airport). The proposed facial embedding is a privacy-aware parameterized function, which maps facial images to high-dimensional vectors in order to facilitate the identification and tracking of individuals. In other words, we are able to uniquely identify person(s) of interest, without the need to determine their true identity. To extract the facial embedding information in crowds, concurrent residual neural network (ResNet) embedding pipeline for each camera is proposed. Specifically, facial embedding feature vectors are generated in real-time by each camera using the proposed enhanced ResNet architecture, which is trained with vectorized-l2-loss function for face recognition. The multivariate kernel density estimation matching algorithm is then applied to facial embedding pipelines generated by cameras at the fog cloud for identity reidentification and security verification. This allows us to ensure the privacy of individuals captured by the camera without compromising on the capability for identity reidentification. Evaluations using mixed datasets in real-time demonstrate that our proposed approach achieves a 2.6% accuracy over other state-of-the-art approaches.