BITS Faculty Publications
Permanent URI for this communityhttp://localhost:4000/handle/123456789/1867
Browse
18 results
Search Results
Item Low-Light Image Enhancement for UAVs With Multi-Feature Fusion Deep Neural Networks(IEEE, 2022-06) Narang, Pratik; Chamola, VinayObject detection in low-light aerial images is a challenging problem due to considerable variation in brightness and varying contrast. Deep learning-based approaches have recently demonstrated great promise in image enhancement. Many existing neural networks used for image quality enhancement first encode the input into low-resolution representations and then decode these representations back to a higher resolution for the contextual information. However, this method leads to the loss of semantic content. Recent research has demonstrated the advantage of maintaining high-resolution information along with lower resolution representations, which maintains image features throughout the network. In this letter, we propose a novel architecture named RNet for low-light image enhancement of aerial images. The proposed network contains multiresolution branches for better understanding of different levels of local and global context through different streams. The performance of RNet is evaluated on a recent synthetic dataset. We also present a comprehensive evaluation with a representative set of state-of-the-art enhancement techniques and neural net architectures.Item Detecting UAV Presence Using Convolution Feature Vectors in Light Gradient Boosting Machine(IEEE, 2022-12) Alladi, Tejasvi; Chamola, VinayThe growing number of Unmanned Aerial Vehicle (UAV) applications brings with it, a rising number of privacy concerns. The high availability of commercial drones is also increasing the need for strict regulations. As far away as we are from establishing such protocols to ensure that the most basic human right to privacy is not exploited, we are further away from enforcing them. Thus, there is a need for a generalised drone detection system to detect different drones operating in a broad range of Radio Frequencies (RF). Previous attempts to tackle this problem have been made using audio, video, radar, WiFi and RF signals. While all these methods have their own benefits and drawbacks, RF has various characteristics which make them suitable for practical applications on a large scale. In this paper, we propose a novel technique called the ConvLGBM model which combines the feature extraction capability of a Convolution Neural Network (CNN) with the high classification accuracy of the Light Gradient Boosting Machine (LightGBM). We develop and evaluate the classifications done by an optimal CNN and the LightGBM model and then compare both models with the ConvLGBM.Item Captionomaly: A Deep Learning Toolbox for Anomaly Captioning in Social Surveillance Systems(IEEE, 2023-01) Chamola, VinayReal-time video stream monitoring is gaining huge attention lately with an effort to fully automate this process. On the other hand, reporting can be a tedious task, requiring manual inspection of several hours of daily clippings. Errors are likely to occur because of the repetitive nature of the task causing mental strain on operators. There is a need for an automated system that is capable of real-time video stream monitoring in social systems and reporting them. In this article, we provide a tool aiming to automate the process of anomaly detection and reporting. We combine anomaly detection and video captioning models to create a pipeline for anomaly reporting in descriptive form. A new set of labels by creating descriptive captions for the videos collected from the UCF-Crime (University of Central Florida-Crime) dataset has been formulated. The anomaly detection model is trained on the UCF-Crime, and the captioning model is trained with the newly created labeled set UCF-Crime video description (UCFC-VD). The tool will be used for performing the combined task of anomaly detection and captioning. Automated anomaly captioning would be useful in the efficient reporting of video surveillance data in different social scenarios. Several testing and evaluation techniques were performed. Source code and dataset: https://github.com/Adit31/Captionomaly-Deep-Learning-Toolbox-for-Anomaly-Captioning.Item Cracking the Anonymous IoT Routing Networks: A Deep Learning Approach(IEEE, 2023-03) Chamola, VinayIn recent years, IoT technology has been one of the most rapidly expanding fields, connecting over 27 billion connected devices worldwide. Increasing security concerns, such as software flaws and cyberattacks, limit the use of IoT devices. Tor, also known as “The Onion Router,” is one of the most popular, secure, and widely deployed anonymous routing systems in IoT networks. Tor is based on a worldwide network of relays operated by volunteers worldwide. Tor continues to be one of the most popular and secure tools against network surveillance, traffic analysis, and information censorship due to its robust use of encryption, authentication, and routing protocols. However, ToR is not anticipated to be entirely safe. The increasing computational capabilities of adversaries threaten Tor's ability to withstand adversarial attacks and maintain anonymity. This article describes the foundation of the Tor network, how it operates, potential attacks against Tor, and the network's defense strategies. In addition, the authors present a framework for deep learning that uses bandwidth performance to identify the server's location in Tor, thereby compromising anonymity. This article examines Tor's network's current and projected future in the Internet of ThingsItem DroneSegNet: Robust Aerial Semantic Segmentation for UAV-Based IoT Applications(IEEE, 2022-04) Narang, Pratik; Chamola, VinayUnmanned Aerial Vehicles (UAVs) are the promising “Flying IoT” devices of the future, which can be equipped with various sensors and cognitive capabilities to perform numerous tasks related to remote sensing, search and rescue operations, object tracking, segmentation of roads and buildings, surveillance, etc. However, these AI-driven tasks require heavy computation and may lead to suboptimal performance with embedded processors on a power-constrained battery-operated drone. This work proposes a novel deep learning approach for performing robust semantic segmentation of aerial scenes captured by UAVs. In our setup, the power-constrained drone is used only for data collection, while the computationally intensive tasks are offloaded to a GPU cloud server. Our architecture performs robust semantic segmentation by learning the segmentation maps from jointly utilizing of aerial scenes along with the respective “elevation maps” in a semi-supervised approach. We propose a three-tier deep learning architecture, wherein the first module aims at preliminary feature extraction from aerial scenes using a backbone feature extractor. The second module captures the spatial dependency between the aerial scenes and their respective elevation maps to obtain better semantic information, which is achieved by a bi-directional LSTM. The third module is aimed at enhancing the performance of semantic segmentation through a semi-supervised approach with an encoder to generate segmentation maps and a decoder to reconstruct feature maps. This semi-supervised feature learning ensures robust extraction along with scalability. The proposed architecture was validated on real-world aerial datasets and achieves state-of-the-art results for aerial image segmentation.Item A deep learning based misbehavior classification scheme for intrusion detection in cooperative intelligent transportation systems(Elsevier, 2022-07) Alladi, Tejasvi; Chamola, VinayWith the rise of the Internet of Vehicles (IoV) and the number of connected vehicles increasing on the roads, Cooperative Intelligent Transportation Systems (C-ITSs) have become an important area of research. As the number of Vehicle to Vehicle (V2V) and Vehicle to Interface (V2I) communication links increases, the amount of data received and processed in the network also increases. In addition, networking interfaces need to be made more secure for which existing cryptography-based security schemes may not be sufficient. Thus, there is a need to augment them with intelligent network intrusion detection techniques. Some machine learning-based intrusion detection and anomaly detection techniques for vehicular networks have been proposed in recent times. However, given the expected large network size, there is a necessity for extensive data processing for use in such anomaly detection methods. Deep learning solutions are lucrative options as they remove the necessity for feature selection. Therefore, with the amount of vehicular network traffic increasing at an unprecedented rate in the C-ITS scenario, the need for deep learning-based techniques is all the more heightened. This work presents three deep learning-based misbehavior classification schemes for intrusion detection in IoV networks using Long Short Term Memory (LSTM) and Convolutional Neural Networks (CNNs). The proposed Deep Learning Classification Engines (DCLE) comprise of single or multi-step classification done by deep learning models that are deployed on the vehicular edge servers. Vehicular data received by the Road Side Units (RSUs) is pre-processed and forwarded to the edge server for classifications following the three classification schemes proposed in this paper. The proposed classifiers identify 18 different vehicular behavior types, the F1-scores ranging from 95.58% to 96.75%, much higher than the existing works. By running the classifiers on testbeds emulating edge servers, the prediction performance and prediction time comparison of the proposed scheme is compared with those of the existing studies.Item Machine-Learning-Assisted Security and Privacy Provisioning for Edge Computing: A Survey(IEEE, 2021-07) Chamola, VinayEdge computing (EC), is a technological game changer that has the ability to connect millions of sensors and provide services at the device end. The broad vision of EC integrates storage, processing, monitoring, and control of operations in the Edge of the network. Though EC provides end-to-end connectivity, speeds up operation, and reduces latency of data transfer, security is a major concern. The tremendous growth in the number of Edge Devices and the amount of sensitive information generated at the device and the cloud creates a broad surface of attack and therefore, the need to secure the static and mobile data is imperative. This article is a comprehensive survey that describes the security and privacy issues in various layers of the EC architecture that result from the networking of heterogeneous devices. Second, it discusses the wide range of machine learning and deep learning algorithms that are applied in EC use cases. Following this, this article broadly details the different types of attacks that the Edge network confronts, and the intrusion detection systems and the corresponding machine learning algorithms that overcome these security and privacy concerns. The details of machine learning and deep learning techniques for EC security are tabulated. Finally, the open issues in securing Edge networks and future research directions are provided.Item DCNN-GA: A Deep Neural Net Architecture for Navigation of UAV in Indoor Environment(IEEE, 2021-03) Chamola, VinayThe applications of unmanned aerial vehicles (UAVs) in military, intelligent transportation, agriculture, rescue operations, natural environment mapping, and many other allied domains has increased exponentially during the past few years. Some of the use cases of their applications range from aerial surveillance, data retrieval to their use in real-time communicative networks. Though UAVs were traditionally used only outdoors, many of its indoor applications like for rescue operations, inventory tracking in warehouses, etc., have recently emerged and these use cases are being actively explored. One of the major challenges for indoor drone applications is navigation and obstacle avoidance. Due to indoor operations, the global positioning system fails in accurate localization and navigation. To address this issue, we introduce a scheme that facilitates the autonomous navigation of UAVs (which have an onboard camera) in the indoor corridors of a building using deep-neural-networks-based processing of images. For a deep neural network, the selection of a good combination of hyperparameters for a better prediction is a complicated task. In this article, the hyperparameters tuning of a convolutional neural network is achieved by using genetic algorithms. The proposed architecture (DCNN-GA) is compared with state-of-the-art ImageNet models. The experimental results show the minimum loss and high performance of the proposed algorithm.Item Role of machine learning and deep learning in securing 5G-driven industrial IoT applications(Elsevier, 2021-12) Chamola, Vinay; Gupta, ShashankThe Internet of Things (IoT) connects millions of computing devices and has set a stage for future technology where industrial use cases like smart cities and smart houses will operate with minimal human intervention. IoT’s cross-domain amalgamations with emergent technologies like 5G and blockchain affects human life. Hence, increase in reliance over IoT necessitates focus on its privacy and security concerns. Implementing security through encryption, authentication, access control and communication security is the need of the hour. These needs can be best catered with the use of machine learning (ML) and deep learning (DL) that can help in realizing secure intelligent systems. In this work, the authors present a comprehensive review for securing Industrial-IoT (I-IoT) devices to contribute to the development of security methods for I-IoT deployed over 5G and blockchain. The survey provides a general analysis of the state-of-the-art security implementation and further assesses the product life cycle of IoT devices. The authors present numerous virtues as well as faults in the machine learning and deep learning algorithms deployed over the fog architecture in context with the security solutions. The potential security algorithms can help overcome many challenges in the IoT security and pave way for implementation with emerging technologies like 5G, blockchain, edge computing, fog computing and their use cases for creating smart environments.