Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Mohsin Shah has a total of 64 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • LiDAR Fusion-based Optimized Path Selection for Autonomous Vehicles using the Verification Model

    by Muhammad Nawaz Khan, Sokjoon Lee, Mohsin Shah, Arshad Khan, Abdul Malik, Ajab Khan

    Autonomous vehicles use LiDAR sensors to navigate 3D maps of their surroundings, enabling improved decision-making and road control systems. It utilizes laser light to measure the speed, distance, shape, direction, and perturbation of any 3D object in a point cloud. Data on point clouds is in high resolution and used to identify different objects of regular or irregular shapes. It is also used in safe path planning, path optimization, and object avoidance routes. Although point cloud data is robust and beneficial for safe navigation, errors and inconsistencies in the interpretation lead to erroneous judgments. Processing wrong information potentially resulting in hazardous circumstances with severe consequences of casualties and road accidents. In contrast, each path planning and path optimization policy should be thoroughly tested and validated before being applied to real-world scenarios. To provide a safer and collision-free path and to verify optimized path policy, we have proposed a technique called ”LiDAR Fusion-based Optimized Path Selection (LOPS) for Autonomous Vehicles using the Verification Model.” The proposed method uses an object detection strategy and instructs autonomous vehicles to use optimized paths for safe travel and collision-free path selection. The proposed scheme is formally modeled and verified for defined states in UPPAAL, and each state with defined properties has been checked. LOPS has been evaluated with different parameters, including the jerking rate, which ranges from 1.0% to 0.06% in various densities of autonomous vehicles. We have also calculated the effect of optimization of the path on the distance covered in unit time, and it is found that 50 vehicles require 20% more time to cover 50 km than 10, 5, or 2 vehicles. The lane usage is calculated in unit time and shows that using optimization in the lane scenario, lane 2 is 5-19 times more used in 5 minutes. LOPS is evaluated with two datasets, CARLA and NuScenes for different parameters, including path length, path smoothness, object avoidance success rate, path tracking accuracy, and decision delay comparisons.

  • Open Access

    A Texture-Aware CNN Predictor for Reversible Data Hiding

    by Mohsin Shah, Chang Choi

    Reversible data hiding (RDH) enables the reversible embedding of additional data into cover media, allowing the cover media to be perfectly recovered after extracting the embedded data. RDH relies on accurate prediction of pixels to generate sharply distributed prediction error histograms, thereby maximizing embedding capacity and minimizing visual distortion. While convolutional neural network (CNN)-based predictors excel in smooth regions of cover images by leveraging local correlation, they often fail to produce accurate predictions in the textured regions. To address the limitation of CNN predictors, we propose a novel attention fusion-based CNN predictor (AFCNNP) that adaptively combines the CNN predictor with a non-local means (NLM) predictor. The proposed fusion framework learns spatial weight maps to favor CNN predictions in smooth regions and NLM predictions in textured regions. The experimental results show that the proposed framework outperforms other state-of-the-art CNN predictors by significantly lowering the mean absolute error, mean squared error, and variance of prediction errors, leading to more accurate pixel predictions. With the proposed fusion framework, the embedding and visual performance of prediction error expansion (PEE)-based RDH is improved compared to typical CNN-based RDH methods.

  • Open Access

    Securing the Internet of Things, Lightweight Mutual Authentication Based on Quantum Key Distribution

    by Muhammad Nawaz Khan, Inam Ullah, Sokjoon Lee, Mohsin Shah

    The Internet of Things (IoT) and quantum computing revolutionized the era of conventional and classical computing into a new paradigm of Quantum-IoT where qubits and entanglement make IoT more interactive, powerful, and secure. They facilitate numerous tasks by increasing productivity and efficiency, paving the path for a smarter and more connected future. In this article, we propose a novel authentication scheme, “Securing the Internet of Things, Lightweight Mutual Authentication Based on Quantum Key Distribution (LMA-QIoT)”. LMA-QIoT enables mutual authentication using various parameters including quantum key distribution, symmetric keys and timestamps, as well as additional quantum random numbers. All these parameters play a crucial role in thwarting man-in-the-middle, backtracking and nonce reuse attacks. The evaluation of LMA-QIoT demonstrates that quantum key distribution and quantum numbers enhance system performance by reducing CPU usage by 25% and memory requirements 30% compared to an IoT edge-based system and without a server, respectively. In the reconfiguration ratio, the efficiency metric grows exponentially and remains constant on the initial line in edge-server-based systems. In comparison, LMA-QIoT confirms a much reduced overall computational complexity by 16.64%, with the lowest computational cost of O(n2). At 1024 Bytes, the original data length and increased data length (normalized) sizes stay constant with 2logn(klogn). Comparing the total overhead, LMA-QIoT demonstrates a reduction of 33 ms, which corresponds to approximately 16.63% less than the baseline mechanisms.


  • Open Access

    LLM-enabled adaptive scheduling in IoT sensing for optimized network performance

    by Muhammad Nawaz Khan, Sokjoon Lee, Sang Su Lee, Mohsin Shah, Inam Ullah, Shakila Basheer, Ali Kashif Bashir

    The use of numerous sensors on edge devices, combined with the emergence of AI techniques, makes the IoT environment more intelligent and interactive. The resulting paradigm encompasses device-centric systems that operate instantly and remotely with zero clicks. However, with these advantages, many functional challenges affect remote sensing, including incomplete data, communication delay, lack of context awareness, and dynamically switching topology. To address these challenges, we have proposed a novel scheme, "LLM-Enabled Adaptive Scheduling in IoT Sensing for Optimized Network Performance (LLM-AS)." This scheme uses LLM to adjust the system's sensing to avoid redundant and useless data sending and enhance decision-making for optimized network resources. First, LLM-AS is trained with a defined data set for different parameters, such as packet loss trends, time-based fluctuations, event triggers, network failure patterns, and congestion signals with contextual decisions. Then, this scheme is deployed in a dynamic remote monitoring system for learning and updating task descriptions to utilize the feedback for future decisions and enhance the system performance. Evaluation of LLM-AS on various parameters using the CASAS dataset shows that the optimization functions of LLM are useful and make the IoT more usable. The LLM-AS optimization function confirms an improvement of 57.8% to 60% in MTP, a decrease of 26% to 60% in median delay, and an optimized energy solution with a confidence interval of 95% and a very small error margin. It also indicates that the precision score is about 0.86, the recall score is about 0.82, and the RMSE is about 0.21; all these values suggest high separability for varying conditions of IoT systems in dynamically changing situations.


  • Open Access

    Exploiting Human Visual Limitations for High Capacity Information Hiding in Digital Images Using Alpha Channel

    by Mohsin Shah, Sokjoon Lee, Muhammad Nawaz Khan, Hyeokchan Kwon

    Information hiding in digital images is a critical technique for secure communication in the era of high-volume multimedia exchange. While many existing methods embed secret data by modifying image parameters in the spatial or frequency domains, they often overlook local pixel characteristics and the potential of the alpha channel for data embedding. This paper proposes a flexible Least Significant Bit (LSB) information hiding framework that adaptively adjusts the number of embedded bits based on pixel intensity ranges. By configuring the embedding function, the framework can support multiple strategies: i) allocating more bits to low-intensity regions and fewer to high-intensity ones, ii) leveraging Weber’s Just Noticeable Difference (JND) principle to embed more data in high-intensity pixels, and iii) combining both approaches to embed more in low- and high-intensity regions while reducing payload in mid-intensity regions. Additionally, the alpha channel is exploited to further enhance embedding capacity while preserving visual fidelity. Experimental results on standard test images demonstrate that the proposed adaptive framework significantly outperforms existing methods in embedding capacity, while maintaining high visual quality with PSNR values consistently above practical thresholds.


  • Open Access

    Frequency domain manipulation of multiple copy-move forgery in digital image forensics

    by Tanzeela Qazi, Mohsin Shah, Mushtaq Ali, Faqir Gul, Muneer Ahmad, Ajmal Khan

    Copy move forgery is a type of image forgery in which a portion of the original image is copied and pasted in a new location on the same image. The consistent illumination and noise pattern make this kind of forgery more difficult to detect. In copy-move forgery detection, conventional approaches are generally effective at identifying simple multiple copy-move forgeries. However, the conventional approaches and deep learning approaches often fall short in detecting multiple forgeries when transformations are applied to the copied regions. Motivated from these findings, a transform domain method for generating and analyzing multiple copy-move forgeries is proposed in this paper. This method utilizes the discrete wavelet transform (DWT) to decompose the original and patch image into approximate (low frequency) and detail coefficients (high frequency). The patch image approximate and details coefficients are inserted into the corresponding positions of the original image wavelet coefficients. The inverse DWT (IDWT) reconstructs the processed image planes after modification which simulates the multiple copy move forgery. In addition, this approach is tested by resizing the region of interest with varying patch sizes resulting in an interesting set of outcomes when evaluated against existing state-of-the-art techniques. This evaluation allows us to identify gaps in existing approaches and suggest improvements for creating more robust detection techniques for multiple copy-move forgeries.


  • Open Access

    Adaptive Scheduling in Cognitive IoT Sensors for Optimizing Network Performance Using Reinforcement Learning

    by Muhammad Nawaz Khan, Sokjoon Lee, Mohsin Shah

    Cognitive sensors are embedded in home appliances and other surrounding devices to create a connected, intelligent environment for providing pervasive and ubiquitous services. These sensors frequently create massive amounts of data with many redundant and repeating bit values. Cognitive sensors are always restricted in resources, and if careful strategy is not applied at the time of deployment, the sensors become disconnected, degrading the system’s performance in terms of energy, reconfiguration, delay, latency, and packet loss. To address these challenges and to establish a connected network, there is always a need for a system to evaluate the contents of detected data values and dynamically switch sensor states based on their function. Here in this article, we propose a reinforcement learning-based mechanism called “Adaptive Scheduling in Cognitive IoT Sensors for Optimizing Network Performance using Reinforcement Learning (ASC-RL)”. For reinforcement learning, the proposed scheme uses three types of parameters: internal parameters (states), environmental parameters (sensing values), and history parameters (energy levels, roles, number of switching states) and derives a function for the state-changing policy. Based on this policy, sensors adjust and adapt to different energy states. These states minimize extensive sensing, reduce costly processing, and lessen frequent communication. The proposed scheme reduces network traffic and optimizes network performance in terms of network energy. The main factors evaluated are joint Gaussian distributions and event correlations, with derived results of signal strengths, noise, prediction accuracy, and energy efficiency with a combined reward score. Through comparative analysis, ASC-RL enhances the overall system’s performance by 3.5% in detection and transition probabilities. The false alarm probabilities are reduced to 25.7%, the transmission success rate is increased by 6.25%, and the energy efficiency and reliability threshold are increased by 35%.


  • Open Access

    An Effective Adaptive Downsampling Method for High-Resolution Multi-Panel Image Segmentation

    by Faqir Gul, Mohsin Shah, Mushtaq Ali, Tanzeela Qazi, Muneer Ahmad, Abid Mehmood

    In modern healthcare, multi-panel images have become increasingly prevalent in medical diagnosis and treatment, representing about 50% of the medical literature. These images integrate diverse imaging modalities, such as X-rays, CT scans, and MRIs, into a unified composite image, facilitating physicians to examine all modalities simultaneously. However, retrieving sub-images from multi-panel images poses a challenge for existing medical image retrieval systems, as they often treat these multi-panel images as single-panel images, thereby restricting access to their constituent sub-images. Consequently, precise segmentation of multi-panel images into sub-images is imperative for effective retrieval. Current segmentation methods are computationally expensive when applied to large-scale, high-resolution multi-panel images.To address this challenge, we propose an adaptive downsampling-based segmentation method that identifies the inter-panel borders separating the sub-images of a multi-panel image by selectively scanning alternate rows and columns, rather than every row and column as done in state-of-the-art approaches.We evaluated the method on a subset of the ImageCLEFmed 2016 dataset, which includes both single-panel and multi-panel images. The experimental results demonstrate that the proposed method significantly reduces computational time while effectively segmenting multi-panel images.