IET Signal Processing最新文献

We propose a multi-objective joint model of non-negative matrix factorization (NMF) and deep neural network (DNN) with a new loss function for speech enhancement. The proposed loss function (L_MOFD) is a weighted combination of a frequency differential spectrum mean squared error (MSE)-based loss function (L_FD) and a multi-objective MSE loss function (L_MO). The conventional MSE loss function computes the discrepancy between the estimated speech and clean speech across all frequencies, disregarding the process of changing amplitude in the frequency domain which contains valuable information. The differential spectrum representation retains spectral peaks that carry important information. Using this representation helps to ensure that this information in the speech signal is reserved. Also, on the other hand, noise spectra typically have a flat shape and as the differential operation makes the flat spectral partly close to zero, the differential spectrum is resistant to noises with smooth structures. Thus, we propose using a frequency-differentiated loss function that considers the magnitude spectrum differentiations between the neighboring frequency bins in each time frame. This approach maintains the spectrum variations of the objective signal in the frequency domain, which can effectively reduce the noise deterioration effects. The multi-objective MSE term (L_MO) is a combined two-loss function related to the NMF coefficients which are the intermediate output targets, and the original spectral signals as the actual output targets. The use of encoded NMF coefficients as low-dimensional structural features for DNN serves as prior knowledge and helps the learning process. L_MO is used beside L_FD to take advantage of both the properties of the original and the differential spectrum in the training loss function. Moreover, a DNN-based noise classification and fusion strategy (NCF) is proposed to exploit a discriminative model for noise reduction. The experiments reveal the improvements of the proposed approach compared to the previous methods.

Compressive sensing (CS) is a technique that enables the recovery of sparse signals using fewer measurements than traditional sampling methods. To address the computational challenges of CS reconstruction, our objective is to develop an interpretable and concise neural network model for reconstructing natural images using CS. We achieve this by mapping one step of the iterative shrinkage thresholding algorithm (ISTA) to a deep network block, representing one iteration of ISTA. To enhance learning ability and incorporate structural diversity, we integrate aggregated residual transformations (ResNeXt) and squeeze-and-excitation mechanisms into the ISTA block. This block serves as a deep equilibrium layer connected to a semi-tensor product network for convenient sampling and providing an initial reconstruction. The resulting model, called MsDC-DEQ-Net, exhibits competitive performance compared to state-of-the-art network-based methods. It significantly reduces storage requirements compared to deep unrolling methods, using only one iteration block instead of multiple iterations. Unlike deep unrolling models, MsDC-DEQ-Net can be iteratively used, gradually improving reconstruction accuracy while considering computation tradeoffs. Additionally, the model benefits from multiscale dilated convolutions, further enhancing performance.

Multipath interference in radar signals caused by sea, ground, and other environments poses significant challenges to the target detection, tracking, and classification capabilities of radar systems. Existing methods for radar signal identification require labeled samples and focus mainly on the classification of normal signals. However, in practice, anomalous samples (multipath interference signals) may be scarce and highly imbalanced (i.e., mostly normal samples). To address this problem, we propose a deep anomaly detection with attention (DADA) for semisupervised detection of multipath radar signals. The method transforms radar signals into time–frequency images and is trained exclusively on normal samples. The autoencoder architecture is extended with a feature extractor network to capture latent sample features. CBAM attention is introduced to improve feature extraction. By learning the distribution of normal samples in high-dimensional image space and low-dimensional feature space, a two-dimensional feature space representing normal samples is constructed. A one-class SVM then learns the boundary of normal samples for anomaly detection. Extensive experiments on radar signal datasets validate the effectiveness of the proposed approach.