CPACT’s Post

Our IEEE COLCOM paper is already available in IEEE Xplore. Enjoy the read. "On the feasibility of using an encoder-only model for anomaly detection: the BERTAD approach.

Hey Connections! Exciting News 🎉 !! I’m thrilled to share that I have presented my research on “Pose Estimation based Fall Detection system using MediaPipe” at ICCSP ’24 10th International Conference on Communication and Signal Processing! The IEEE conference presentation provided me with an amazing opportunity to network with other research scholars and industry experts and present my work. I was able to gain relevant insights in various verticals of cutting edge technology, used in diverse domains. I’m delighted to share that my presented research work has been accepted into the prestigious IEEE Xplore Digital Library: https://lnkd.in/gBD_QSW6 The research focuses on developing a lightweight Fall Detection system which is able to run inference on CPU and edge devices, having decent FPS and accuracy. A Vision-centric approach was considered by leveraging the Pose Estimation technology and skeleton extraction, giving us the body-joint (landmarks) coordinates. A popular lightweight pose estimator module, MediaPipe has been used to extract these pose landmarks. These landmarks are used within a decision condition based system, which reads input frame by frame from the video sequences and detects if any falling action frames are present in the video sequence. The main highlights of my research work includes : - Tackling the problem of Fall Detection using a Computer Vision-centric approach - Leveraging Pose Estimation technology in Computer Vision to detect falling action frames within RGB video sequences - Developing a Decision condition based system (using the Pose landmarks) to detect falls, which is different from existing deep learning based methodologies usually used for Fall Detection in video sequences - Utilizing a lightweight Pose Estimator module, MediaPipe, which is able to run inference on CPU and edge devices with decent FPS and accuracy to detect falls - Making an efficient, lightweight and robust solution to detect falls in video sequences which can be easily integrated into healthcare and biomedical domains for elderly supervision in future, subjected to potential advancements and improvements It is my pleasure to express with a deep sense of gratitude to my mentor, Dr. Malathi G. for her constant insightful guidance and invaluable support throughout this research work. Special mention to my close friends Akshat Dhamale and Krishna Sarawagi for their motivational support and helpful nature, helping me along my journey! And lastly, I want to acknowledge everyone who contributed to this achievement :) Looking forward to more research opportunities! 🙂 #ICCSP2024 #IEEE #ComputerVision #AIinHealthcare #ResearchPresentation #ResearchPaper #MediaPipe

Our latest publication in a Q1 journal (South Korean): Amira Eleff , Mohamed Hamed, Hamed Nassar, "Effectiveness of guard zone in mitigating interference in D2D underlaid cellular networks," Journal of Wireless Mobile Networks, Ubiquitous Computing, and Dependable Applications, Vol. 14, No. 3, pp. 112--124, 2023. https://lnkd.in/drwrTXXs In this article we leverage stochastic geometry to characterize the interference inflicted on cellular users by D2D communications underlaying the cellular channels, and evaluate how a guard zone can be a remedy.

I am excited to share that our research article: "FD MU-MIMO Systems: Performance Analysis in the Presence of Imperfect CSI and Non-ideal Transceivers" has been accepted for publication in IEEE Transactions on Mobile Computing https://lnkd.in/gjEKiu4T

Happy to share that our paper “Joint Power and User Allocation in Coexistence of eMBB and URLLC Services” has been accepted for publication in IEEE Communications Letters. In this paper, the coexistence of enhanced mobile broadband (eMBB) and ultra-reliable low latency communication (URLLC) services in the intelligent reconfigurable surface (IRS)-assisted terahertz (THz) network is analyzed. The coexistence of eMBB and URLLC services in the same network leads to a challenging resource management problem. We formulate the joint power and user allocation (JPUA) to maximize the achievable data rate of eMBB users. Specifically, we propose a one-to-many matching game to allocate the IRSs to the eMBB users. The simulation outcomes indicate that the proposed scheme outperforms the baseline methods regarding the sum data rate for eMBB users. Digital Object Identifier: 10.1109/LCOMM.2024.3411399 https://lnkd.in/eaFCR9Sa I'm truly grateful for the invaluable guidance and support provided by my distinguished co-authors, Prof. Georges Kaddoum, Prof. Tri-Nhu Do, and Thanh Luan Nguyen.

I'm thrilled to share that our latest paper, "Steady-State-Aware Model Predictive Control for Tracking in Systems With Limited Computing Capacity," has been successfully published in IEEE Control Systems Letters! Special thanks to Dr. Mehdi Hosseinzadeh for his invaluable support and contribution. We've developed a novel model predictive control capable of addressing both transient and steady-state specifications in systems with limited computing power. Keen to delve deeper? Discover the details of our research by clicking on the link below: https://lnkd.in/g7vqrK3A #ModelPredictiveControl #ComputingCapacity #Steady_State_Configuration

This paper proposes #Model-#based #Data #Augmentation #for #Deep #learning-#based #Modulation #Classification (#MDA-#DMC), to build a #high-#quality #dataset from a #small #amount #of #labeled #seed #data. MDA-DMC leverages two well-known augmentation methods: #adding #Gaussian #noise to, and #rotation #of #the #seed #signal #constellations. Furthermore, the authors develop two #novel #augmentation #methods to combat channel and hardware impairments: #radial #shift and #stretching #of #the #signal #constellations. They are the first to investigate the correlation between these augmentation methods and the #channel/ #hardware #impairments, demonstrating the adverse effect of the rotation and stretching of signal constellations on classifier performance.----Erma Perenda, Gérôme Bovet, Mariya Zheleva, Sofie Pollin More details can be found at this link: https://lnkd.in/gvQ5uyU8

This letter examines the capacity of long-range (#LoRa) networks using #stochastic #geometry and #queuing #theory. The authors take into account limits on how often devices can transmit, as well as issues that arise when different #spreading #factors (SFs) interfere with each other. The letter measures capacity by looking at both the maximum number of devices per area and how much data each device sends. They introduce #cross-#correlation #factors to quantify problems caused by SFs not being completely independent. ---- Yathreb Bouazizi More details can be found at this link: https://lnkd.in/gvSB2nm7

Excited to share our latest publication, "Optimizing Cooperative Perception Message Dissemination in Vehicular Networks Through Selective Transmission," presented at the 16th International Conference on COMmunication Systems & NETworkS (COMSNETS) 2024. We've addressed the challenge of vehicle selection for the transmission of Collective Perception Messages (CPMs) in Cellular Vehicle-to-Everything (C-V2X). Our objective was to minimize network congestion without significantly affecting perception coverage. We formulated this problem as an Integer Linear Programming (ILP) problem and suggested a solution using an approximation algorithm. The outcome of our research reveals a considerable decrease in network load and an improvement in the efficiency of CPM distribution. A big thanks to my co-authors Rishav Singh for his hard work and Antony Franklin for the guidance. Take a closer look at our research and let us know your thoughts. The full paper is available here: https://lnkd.in/gfrMPXpF #CV2X #5G #VehicularNetworks #CPM #RoadSafety #ILP #ApproximationAlgorithm #COMSNETS2024