arXiv subjects

Khuram Naveed

Publications and source records attributed to Khuram Naveed.

11 recordsLinked to original sources

arXiv2026

HARU-Net: Hybrid Attention Residual U-Net for Edge-Preserving Denoising in Cone-Beam Computed Tomography

Cone-beam computed tomography (CBCT) is widely used in dental and maxillofacial imaging, but low-dose acquisition introduces strong, spatially varying noise that degrades soft-tissue visibility and obscures fine anatomical structures. Classical denoising methods struggle to suppress noise in CBCT while preserving edges. Although deep learning-based approaches offer high-fidelity restoration, their use in CBCT denoising is limited by the scarcity of high-resolution CBCT data for supervised training. To address this research gap, we propose a novel Hybrid Attention Residual U-Net (HARU-Net) for high-quality denoising of CBCT data, trained on a cadaver dataset of human hemimandibles acquired using a high-resolution protocol of the 3D Accuitomo 170 (J. Morita, Kyoto, Japan) CBCT system. The novel contribution of this approach is the integration of three complementary architectural components: (i) a hybrid attention transformer block (HAB) embedded within each skip connection to selectively emphasize salient anatomical features, (ii) a residual hybrid attention transformer group (RHAG) at the bottleneck to strengthen global contextual modeling and long-range feature interactions, and (iii) residual learning convolutional blocks to facilitate deeper, more stable feature extraction throughout the network. HARU-Net consistently outperforms state-of-the-art (SOTA) methods including SwinIR and Uformer, achieving the highest PSNR (37.52 dB), highest SSIM (0.9557), and lowest GMSD (0.1084). This effective and clinically reliable CBCT denoising is achieved at a computational cost significantly lower than that of the SOTA methods, offering a practical advancement toward improving diagnostic quality in low-dose CBCT imaging.

eess.IV↗

arXiv2025

A Fully Multivariate Multifractal Detrended Fluctuation Analysis Method for Fault Diagnosis

We propose a fully multivariate generalization of multifractal detrended fluctuation analysis (MFDFA) and leverage it to develop a fault diagnosis framework for multichannel machine vibration data. We introduce a novel covariance-weighted $L_{pq}$ matrix norm based on Mahalanobis distance to define a fully multivariate fluctuation function that uniquely captures cross-channel dependencies and variance biases in multichannel vibration data. This formulation, termed FM-MFDFA, allows for a more accurate characterization of the multiscale structure of multivariate signals. To enhance feature relevance, the proposed framework integrates multivariate variational mode decomposition (MVMD) to isolate fault-relevant components before applying FM-MFDFA. Results on wind turbine gearbox data demonstrate that the proposed method outperforms conventional MFDFA approaches by effectively distinguishing between healthy and faulty machine states, even under noisy conditions.

eess.SP↗

arXiv2025

A Contrastive Learning Framework for Breast Cancer Detection

Breast cancer, the second leading cause of cancer-related deaths globally, accounts for a quarter of all cancer cases [1]. To lower this death rate, it is crucial to detect tumors early, as early-stage detection significantly improves treatment outcomes. Advances in non-invasive imaging techniques have made early detection possible through computer-aided detection (CAD) systems which rely on traditional image analysis to identify malignancies. However, there is a growing shift towards deep learning methods due to their superior effectiveness. Despite their potential, deep learning methods often struggle with accuracy due to the limited availability of large-labeled datasets for training. To address this issue, our study introduces a Contrastive Learning (CL) framework, which excels with smaller labeled datasets. In this regard, we train Resnet-50 in semi supervised CL approach using similarity index on a large amount of unlabeled mammogram data. In this regard, we use various augmentation and transformations which help improve the performance of our approach. Finally, we tune our model on a small set of labelled data that outperforms the existing state of the art. Specifically, we observed a 96.7% accuracy in detecting breast cancer on benchmark datasets INbreast and MIAS.

cs.CV↗

arXiv2025

InJecteD: Analyzing Trajectories and Drift Dynamics in Denoising Diffusion Probabilistic Models for 2D Point Cloud Generation

This work introduces InJecteD, a framework for interpreting Denoising Diffusion Probabilistic Models (DDPMs) by analyzing sample trajectories during the denoising process of 2D point cloud generation. We apply this framework to three datasets from the Datasaurus Dozen bullseye, dino, and circle using a simplified DDPM architecture with customizable input and time embeddings. Our approach quantifies trajectory properties, including displacement, velocity, clustering, and drift field dynamics, using statistical metrics such as Wasserstein distance and cosine similarity. By enhancing model transparency, InJecteD supports human AI collaboration by enabling practitioners to debug and refine generative models. Experiments reveal distinct denoising phases: initial noise exploration, rapid shape formation, and final refinement, with dataset-specific behaviors example, bullseyes concentric convergence vs. dinos complex contour formation. We evaluate four model configurations, varying embeddings and noise schedules, demonstrating that Fourier based embeddings improve trajectory stability and reconstruction quality

cs.LG↗

arXiv2025

Impact of Labeling Inaccuracy and Image Noise on Tooth Segmentation in Panoramic Radiographs using Federated, Centralized and Local Learning

Objectives: Federated learning (FL) may mitigate privacy constraints, heterogeneous data quality, and inconsistent labeling in dental diagnostic AI. We compared FL with centralized (CL) and local learning (LL) for tooth segmentation in panoramic radiographs across multiple data corruption scenarios. Methods: An Attention U-Net was trained on 2066 radiographs from six institutions across four settings: baseline (unaltered data); label manipulation (dilated/missing annotations); image-quality manipulation (additive Gaussian noise); and exclusion of a faulty client with corrupted data. FL was implemented via the Flower AI framework. Per-client training- and validation-loss trajectories were monitored for anomaly detection and a set of metrics (Dice, IoU, HD, HD95 and ASSD) was evaluated on a hold-out test set. From these metrics significance results were reported through Wilcoxon signed-rank test. CL and LL served as comparators. Results: Baseline: FL achieved a median Dice of 0.94889 (ASSD: 1.33229), slightly better than CL at 0.94706 (ASSD: 1.37074) and LL at 0.93557-0.94026 (ASSD: 1.51910-1.69777). Label manipulation: FL maintained the best median Dice score at 0.94884 (ASSD: 1.46487) versus CL's 0.94183 (ASSD: 1.75738) and LL's 0.93003-0.94026 (ASSD: 1.51910-2.11462). Image noise: FL led with Dice at 0.94853 (ASSD: 1.31088); CL scored 0.94787 (ASSD: 1.36131); LL ranged from 0.93179-0.94026 (ASSD: 1.51910-1.77350). Faulty-client exclusion: FL reached Dice at 0.94790 (ASSD: 1.33113) better than CL's 0.94550 (ASSD: 1.39318). Loss-curve monitoring reliably flagged the corrupted site. Conclusions: FL matches or exceeds CL and outperforms LL across corruption scenarios while preserving privacy. Per-client loss trajectories provide an effective anomaly-detection mechanism and support FL as a practical, privacy-preserving approach for scalable clinical AI deployment.

eess.IV↗

arXiv2025

NAADA: A Noise-Aware Attention Denoising Autoencoder for Dental Panoramic Radiographs

Convolutional denoising autoencoders (DAEs) are powerful tools for image restoration. However, they inherit a key limitation of convolutional neural networks (CNNs): they tend to recover low-frequency features, such as smooth regions, more effectively than high-frequency details. This leads to the loss of fine details, which is particularly problematic in dental radiographs where preserving subtle anatomical structures is crucial. While self-attention mechanisms can help mitigate this issue by emphasizing important features, conventional attention methods often prioritize features corresponding to cleaner regions and may overlook those obscured by noise. To address this limitation, we propose a noise-aware self-attention method, which allows the model to effectively focus on and recover key features even within noisy regions. Building on this approach, we introduce the noise-aware attention-enhanced denoising autoencoder (NAADA) network for enhancing noisy panoramic dental radiographs. Compared with the recent state of the art (and much heavier) methods like Uformer, MResDNN etc., our method improves the reconstruction of fine details, ensuring better image quality and diagnostic accuracy.

eess.IV↗

arXiv2023

EDDense-Net: Fully Dense Encoder Decoder Network for Joint Segmentation of Optic Cup and Disc

Glaucoma is an eye disease that causes damage to the optic nerve, which can lead to visual loss and permanent blindness. Early glaucoma detection is therefore critical in order to avoid permanent blindness. The estimation of the cup-to-disc ratio (CDR) during an examination of the optical disc (OD) is used for the diagnosis of glaucoma. In this paper, we present the EDDense-Net segmentation network for the joint segmentation of OC and OD. The encoder and decoder in this network are made up of dense blocks with a grouped convolutional layer in each block, allowing the network to acquire and convey spatial information from the image while simultaneously reducing the network's complexity. To reduce spatial information loss, the optimal number of filters in all convolution layers were utilised. In semantic segmentation, dice pixel classification is employed in the decoder to alleviate the problem of class imbalance. The proposed network was evaluated on two publicly available datasets where it outperformed existing state-of-the-art methods in terms of accuracy and efficiency. For the diagnosis and analysis of glaucoma, this method can be used as a second opinion system to assist medical ophthalmologists.

eess.IV↗

arXiv2023

A Tutorial on Modeling and Control of Slippage in Wheeled Mobile Robots

This tutorial intends to present a deep insight of the slip and skid problem in WMRs. Specifically, we present a brief introduction to slippage and the limitations it imposes on WMR's applications. In addition, we shed some light on state of the art review on modeling methodologies and controlling mechanisms for slip and skid in WMRs. This paper also discusses the future trends in the use of WMRs and possibility of new domains in their use in modern world.

eess.SP↗

arXiv2023

Multivariate Signal Denoising Based on Generic Multivariate Detrended Fluctuation Analysis

We propose a novel multivariate signal denoising method that performs long-range correlation analysis of multiple modes in input data by considering inherent inter-channel dependencies of the data. That is achieved through a novel and generic multivariate extension of detrended fluctuation analysis (DFA) method - another contribution of this paper. Specifically, our proposed denoising method first obtains data driven multiscale signal representation using multivariate variational mode decomposition (MVMD) method. Then, the proposed generic multivariate DFA is used to reject the predominantly noisy modes based on their randomness scores. Finally, the denoised signal is reconstructed by summing the remaining modes albeit after the removal of the noise traces using the principal component analysis (PCA).

eess.SP↗

arXiv2020

A Statistical Approach to Signal Denoising Based on Data-driven Multiscale Representation

We develop a data-driven approach for signal denoising that utilizes variational mode decomposition (VMD) algorithm and Cramer Von Misses (CVM) statistic. In comparison with the classical empirical mode decomposition (EMD), VMD enjoys superior mathematical and theoretical framework that makes it robust to noise and mode mixing. These desirable properties of VMD materialize in segregation of a major part of noise into a few final modes while majority of the signal content is distributed among the earlier ones. To exploit this representation for denoising purpose, we propose to estimate the distribution of noise from the predominantly noisy modes and then use it to detect and reject noise from the remaining modes. The proposed approach first selects the predominantly noisy modes using the CVM measure of statistical distance. Next, CVM statistic is used locally on the remaining modes to test how closely the modes fit the estimated noise distribution; the modes that yield closer fit to the noise distribution are rejected (set to zero). Extensive experiments demonstrate the superiority of the proposed method as compared to the state of the art in signal denoising and underscore its utility in practical applications where noise distribution is not known a priori.

eess.SP↗

arXiv2020

Wavelet based multivariate signal denoising using Mahalanobis distance and EDF statistics

A multivariate signal denoising method is proposed which employs a novel multivariate goodness of fit (GoF) test that is applied at multiple data scales obtained from discrete wavelet transform (DWT). In the proposed multivariate GoF test, we first utilize squared Mahalanobis distance (MD) measure to transform input multivariate data residing in M-dimensional space $\mathcal{R}^M$ to a single-dimensional space of positive real numbers $\mathcal{R}_+$, i.e., $\mathcal{R}^M \rightarrow \mathcal{R}_+$, where $M > 1$. Owing to the properties of the MD measure, the transformed data in $\mathcal{R}_+$ follows a distinct distribution. That enables us to apply the GoF test using statistic based on empirical distribution function (EDF) on the resulting data in order to define a test for multivariate normality. We further propose to apply the above test locally on multiple input data scales obtained from discrete wavelet transform, resulting in a multivariate signal denoising framework. Within the proposed method, the reference cumulative distribution function (CDF) is defined as a quadratic transformation of multivariate Gaussian random process. Consequently, the proposed method checks whether a set of DWT coefficients belong to multivariate reference distribution or not; the coefficients belonging to the reference distribution are discarded. The effectiveness of our proposed method is demonstrated by performing extensive simulations on both synthetic and real world datasets.

eess.SP↗