SIPL Projects

We address the problem of multichannel audio signal enhancement in reverberant environments with interfering sources. We propose an approach that leverages the Riemannian geometry of the spatial correlation matrices of the received signals to estimate the relative transfer function (RTF) of the desired source. Specifically, we compute the spatial correlation matrices in short-time segments, and subsequently, their Riemannian mean, which preserves shared spectral components while attenuating unshared ones. This enables an effective intermittent interference rejection, leading to accurate RTF estimation.

Nowadays, the metric used to calculate the statistic distance between different datasets is the FID (Frchet Inception Distance): it uses a pretrained inception network and a divergence very close to the W2 divergence (in this case) to approach the distance between them. We assume that the latent representation of each dataset has a Gaussian distribution. We also assume that the Gaussian distribution is not degenerate: we assume that the covariance matrix is a positive definite matrix: all the principal components are not zeros. To these assumptions, we can add the numerical instability and the impossibility to score a single image.

The direction of Arrival (DOA) estimation has been widely applied in radar, sonar, and wireless communications. It refers to the process of retrieving the direction information of several signals from the outputs of several receiving antennas that form a sensor array. Various DOA estimation algorithms have been proposed for this purpose. Such as MUSIC and Beamforming, these are one of the recognized algorithms for DOA estimation that have a good performance when applied on sources that are far enough from the antennas (far field). These algorithms and most of the DOA estimation algorithms use the far field assumption in order to work as intended.

When sampling an audio signal and converting it to a digital, discrete signal, we face several challenges that put the samples at risk of loss of information. One of those risks is driven by the dynamic range of the sampler. A signal that crosses the dynamic range is clipped, and therefore its reconstruction is damaged. In the following discussion, we are about to introduce analyzation results of Beyond Bandwidth Residual Reconstruction (22) algorithm implementation. Here we propose a robust, real-time algorithm that is designed to sample signals that cross the dynamic range with minimum loss of information, relying on an input signal that had been activated by a modulo operator.

In this project, we describe the phase transition of the homological connectivity of a random complex, constructed on a bounded manifold, which was chosen to be a flat cylinder of a general dimension. We do this by generating a random cloud on the cylinder, on which we construct a filtration of a simplicial complex, chosen to be the ech Complex. We try to describe the phase transition of the homological connectivity of the complex i.e., the transition in which the homology of the random complex becomes identical to that of the underlying cylinder. We expect the addition of the boundary (which isn't trivial) to introduce a significant change to the threshold.

The goal of this work is learning the super-resolution space, which is one of the challenges presented in NTIRE 2021 competition. The mission was to solve the challenge according to the competition rules - given low resolution image, to produce good quality super-resolution image. In reality, many high-resolution images can be downsampled to the same low-resolution image. The challenge focused on producing arbitrary number of super-resolution images capturing meaningful diversity, using the same input of low-resolution image. in addition, the output images need to be consistent to the input, and with high photo-realism as perceived by humans.

High precision indoor position estimation enables new opportunities for a variety of commercial, industrial and consumer applications. In this work, we consider a phase-based method to calculate range from noisy measurements of a Frequency Comb in a multi-fading environment. It can be used to determine the range between devices for the next-generation High Accuracy Distance Measurement (HADM) protocol. We have conducted a quantitative analysis of various estimation approaches, considering both Monte-Carlo simulations of synthetic data in a variety of ranges.

The goal of this project was to perform Proof of Concept (PoC) of a theory developed in the SIPL lab by Prof. Ronen Talmon and Or Yair. The theory describes an algorithm that is based on Manifold Learning tools, namely Diffusion Maps, and is used to analyze physical systems and their dynamics empirically meaning without any prior knowledge. The theory was previously only tested on simulated data, so our first and main objective was to perform the PoC by running the algorithm on data acquired from a real physical system. For simplicity, we chose the physical pendulum as our test system.

...

...

...

...

...

...

...

...