4 December 2020 in IEEE Journal of Selected Topics in Signal Processing
Dynamic range limitations in signal processing often lead to clipping, or saturation, in signals. The task of audio declipping is estimating the original audio signal, given its clipped measurements, and has attracted much interest in recent years. Audio declipping algorithms often make assumptions about the underlying signal, such as sparsity or low-rankness, and about the measurement system. In this paper, we provide an extensive review of audio declipping algorithms proposed in the literature. For each algorithm, we present assumptions that are made about the audio signal, the modeling domain, and the optimization algorithm. Furthermore, we provide an extensive numerical evaluation of popular declipping algorithms, on real audio data. We evaluate each algorithm in terms of the Signal-to-Distortion Ratio, and also using perceptual metrics of sound quality. The article is accompanied by a repository containing the evaluated methods.
8 April 2025 in IEEE Access
This paper presents an advanced perimeter protection system that integrates phase-sensitive Optical Time-Domain Reflectometry (Φ-OTDR) with Convolutional Neural Networks (CNNs) for real-time event classification near optical fibers. The proposed approach enhances traditional security methods by providing robust monitoring in challenging environments, such as low visibility and large-scale areas. We evaluated multiple signal preprocessing techniques, including Fast Fourier Transform (FFT), Redundant Discrete Fourier Transform (RDFT), Discrete Wavelet Transform (DWT), and Mel-Frequency Cepstral Coefficients (MFCC), to optimize classification accuracy and computational efficiency. While MFCC achieved the highest accuracy (85.61 %), RDFT provided the best balance between performance (85.47 %) and real-time feasibility, making it the preferred method for deployment. The system successfully differentiates events such as vehicle movement, fence manipulation, and construction work, while anomaly detection capabilities further enhance security by identifying irregular activities with minimal error. These findings demonstrate the potential of integrating fiber-optic sensing with deep learning to develop scalable, real-time perimeter protection solutions for critical infrastructure, border surveillance, and urban security.