FAFauzia Ahmad

Fauzia Ahmad

*This is an algorithmically generated author record based on publicly available information.

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


Fauzia Ahmad has a total of 43 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Direction‐of‐arrival estimation via coarray‐domain RELAX algorithm with source number estimation

    by Fauzia Ahmad, Moeness G. Amin

    A modified RELAX algorithm based on iterative coarray‐domain beamforming for fast source direction‐of‐arrival (DOA) estimation with fully augmentable sparse arrays is proposed. The authors exploit the deterministic centralised nature of the noise in the coarray domain and the Hermitian symmetry of the spatial autocorrelation function to efficiently incorporate source number estimation within the iterative framework. In doing so, the proposed algorithm allows low‐complexity, fast DOA estimation of more sources than sensors, without resorting to computationally expensive implementations of source number estimation using information theoretic criteria. Three variants of the proposed algorithm are presented, each differing in terms of the specific method employed for source number estimation. Extensive simulations are performed with a minimum redundancy array to compare and contrast the performance of the three variants in terms of their accuracy in estimating the number of sources in the field‐of‐view of the array. The results demonstrate that the modified RELAX algorithm can provide accurate estimates of the number and directions of sources, especially when the number of uncorrelated sources is equal to or higher than the number of sensors.


  • Open Access

    Whitening-Aided Learning from Radar Micro-Doppler Signatures for Human Activity Recognition

    by Zahra Sadeghi Adl, Fauzia Ahmad

    Deep learning architectures are being increasingly adopted for human activity recognition using radar technology. A majority of these architectures are based on convolutional neural networks (CNNs) and accept radar micro-Doppler signatures as input. The state-of-the-art CNN-based models employ batch normalization (BN) to optimize network training and improve generalization. In this paper, we present whitening-aided CNN models for classifying human activities with radar sensors. We replace BN layers in a CNN model with whitening layers, which is shown to improve the model’s accuracy by not only centering and scaling activations, similar to BN, but also decorrelating them. We also exploit the rotational freedom afforded by whitening matrices to align the whitened activations in the latent space with the corresponding activity classes. Using real data measurements of six different activities, we show that whitening provides superior performance over BN in terms of classification accuracy for a CNN-based classifier. This demonstrates the potential of whitening-aided CNN models to provide enhanced human activity recognition with radar sensors.


  • Convolutional denoising of underwater laser line scan images via transfer learning

    by Christopher Campbell, Fauzia Ahmad

    In this paper, we consider the problem of Poisson noise suppression in underwater laser line scan (LLS) imagery for image quality enhancement. We investigate two different denoising neural network architectures, one based on a convolutional autoencoder (CAE) and a commonly used convolutional neural network (CNN) based denoiser. Due to the relative abundance of camera images over underwater LLS imagery, we employ transfer learning with camera images as training data for the CAE and CNN architectures. Poisson noise is introduced in these training images at varying levels to mimic a noise-dominated LLS system. Images from an underwater scene under different turbidity conditions are used for testing. Using a composite loss function consisting of l1 and l2 norms, we demonstrate the noise suppression capabilities of both architectures in underwater LLS images via transfer learning. The results show that the CAE outperforms the CNN denoiser qualitatively and quantitatively in terms of the contrast ratio and contrast signal-to-noise ratio.

  • Open Access

    Laboratory and Field Experimental Study of Underwater Inflatable Co-prime Sonar Array (UICSA)

    by Yanjun Li, Bing Ouyang, Tongdi Zhou, Jordan Thomas, Shadi Bavar, Ethan Weber, Lorenzo Michieletto, Tsung-Chow Su, Fauzia Ahmad

    This paper discusses the design and initial testing of a novel hydrophone array system dubbed the Underwater Inflatable Co-prime Sonar Array (UICSA). The UICSA will be a crucial component of an underwater deployable sensing network that can be rapidly deployed using compact autonomous underwater vehicles (AUVs). The UICSA initially is packed in a compact container to fit the payload space of an AUV. After deployment, the UICSA expands to its predetermined full length to acquire sensing data for source localization. More specifically, the mechanical compression of the UICSA is achieved through a non-rigid array support structure, which consists of flexible inflatable segments between adjoining hydrophones that are folded in order to package the UICSA for deployment. The system exploits compression in hydrophone layouts by utilizing a sparse array configuration, namely the co-prime array since it requires fewer hydrophones than a uniform linear array of the same length to estimate a given number of sources. With two-way compression, the storage, handling, and transportation of the compactly designed UICSA is convenient, particularly for the AUVs with limited payload space. The deployment concept and process are discussed, as well as the various UICSA designs of different support structures are described. A comparison of the various mechanical designs is presented and a novel hybrid-based expansion prototype is documented in detail. Laboratory study results of the UICSA prototype are presented that include water-swollen material tests in a pressurized environment and water tank validation of the inflation process. The UICSA prototype also has been deployed in the Harbor Branch channel to validate the performance, the related field test details and source localization results.


  • Robust Copula-Based Detection of Shallow-Buried Landmines With Forward-Looking Radar

    by Afief Dias Pambudi, Fauzia Ahmad, Abdelhak M. Zoubir

    We propose a technique for landmine detection using forward-looking ground penetrating radar. The detector is applied to radar images obtained from multiple viewpoints of the region of interest and is based on a robust version of the likelihood ratio test (LRT). We incorporate the statistical dependence between multi-view images into the test via copula-based model. The test is designed to maximize the worst case performance over all feasible pairs of target and clutter distributions, thereby eliminating the need for strong assumptions about the image statistics. We evaluate the detection performance of the proposed technique for different copula functions representing the dependence structure. Using electromagnetic modeled data of shallow-buried targets under varying ground surface roughness profiles, we demonstrate the superiority of the robust copula-based detector over existing parametric and robust LRT approaches designed under the assumption of statistical independence of multi-view images.


  • Open Access

    Numerical Study of Underwater Inflatable Co-Prime Sonar Array

    by Yanjun Li, Jordan Thomas, Bing Ouyang, Tsung-Chow Su, Fauzia Ahmad

  • Open Access

    Numerical Study of Underwater Inflatable Co-Prime Sonar Array

    by Yanjun Li, Jordan Thomas, Bing Ouyang, Tsung-Chow Su, Fauzia Ahmad

    Underwater Inflatable Co-prime Sonar Array (UICSA) is a compact sonar array assembly which can be deployed in the ocean then morph into predetermined length to work. UICSA design is based on two-way compression concept: algorithm compression and mechanical compression. The algorithm compression employs compressive sensing techniques to enable operation with a sparse co-prime