On the Use of Autoregressive Methods for Audio Inpainting

The paper presents an evaluation of popular audio inpainting methods based on autoregressive modelling, namely, extrapolation-based and Janssen methods. A novel variant of the Janssen method suitable for gap inpainting is also proposed. The main differences between the particular popular approaches are pointed out, and a mid-scale computational experiment is presented. The results demonstrate the importance of the choice of the AR model estimator and the suitability of the new gap-wise Janssen method.

Improving DCE-MRI through unfolded low-rank + sparse optimisation

A method for perfusion imaging with DCE-MRI is developed based on two popular paradigms: the low-rank + sparse model for optimisation-based reconstruction, and the deep unfolding. A learnable algorithm derived from a proximal algorithm is designed with emphasis on simplicity and interpretability. The resulting deep network is trained and evaluated using a simulated measurement of a rat with a brain tumor, showing large performance gain over the classical low-rank + sparse baseline. Moreover, quantitative perfusion analysis is performed based on the reconstructed sequence, proving that even training based on a simple pixel-wise error can lead to significant improvement of the quality of the perfusion maps.

Multiple Hankel matrix rank minimization for audio inpainting

Sasaki et al. (2018) presented an efficient audio declipping algorithm, based on the properties of Hankel-structured matrices constructed from time-domain signal blocks. We adapt their approach to solve the audio inpainting problem, where samples are missing in the signal. We analyze the algorithm and provide modifications, some of them leading to an improved performance. Overall, it turns out that the new algorithms perform reasonably well for speech signals but they are not competitive in the case of music signals.

Analysis Social Sparsity Audio Declipper

We develop the analysis (cosparse) variant of the popular audio declipping algorithm of Siedenburg et al. Furthermore, we extend it by the possibility of weighting the time-frequency coefficients. We examine the audio reconstruction performance of several combinations of weights and shrinkage operators. We show that weights improve the reconstruction quality in some cases; however, the overall scores achieved by the non-weighted are not surpassed. Yet, the analysis Empirical Wiener (EW) shrinkage was able to reach the quality of a computationally more expensive competitor, the Persistent Empirical Wiener (PEW). Moreover, the proposed analysis variant using PEW slightly outperforms the synthesis counterpart in terms of an auditory-motivated metric.

Audio declipping performance enhancement via crossfading

Some audio declipping methods produce waveforms that do not fully respect the physical process of clipping, which is why we refer to them as inconsistent. This letter reports what effect on perception it has if the solution by inconsistent methods is forced consistent by postprocessing. We first propose a simple sample replacement method, then we identify its main weaknesses and propose an improved variant. The experiments show that the vast majority of inconsistent declipping methods significantly benefit from the proposed approach in terms of objective perceptual metrics. In particular, we show that the SS PEW method based on social sparsity combined with the proposed method performs comparable to top methods from the consistent class, but at a computational cost of one order of magnitude lower.