Marco Crocco

A Bilinear Approach to the Position Self-Calibration of Multiple Sensors

This paper presents a novel algorithm for the automatic 3D localization of a set of sensors in an unknown environment. Given the measures of a set of time of arrival delays at each sensor, the approach simultaneously estimates the 3D position of the sensors and the sources that have generated the event. Such inference is obtained with no assumption about the sensor localization; the only assumption made is that the emission time of the sources must be known in order to evaluate the time of flight for each event. Moreover, we propose a further method that deals with the likely case of missing data in the measurements. This occurs when sensors are far apart or behind natural barriers that avoids the registration of the given event. Simulated experiments show the validity of the approach for different setups of sensors and number of events.

Design of Robust Superdirective Arrays With a Tunable Tradeoff Between Directivity and Frequency-Invariance

Frequency-invariant beam patterns are often required by systems using an array of sensors to process broadband signals. If the spatial aperture is shorter than the involved wavelengths, using a superdirective beamforming is essential to get an efficient system. In this context, robustness of array imperfections is a crucial feature. In the literature, only a few approaches have been proposed to design a robust, superdirective, frequency-invariant beamformer based on filter-and-sum architecture; all of them achieve frequency-invariance by imposing a desired beam pattern. However, the choice of a suitable desired beam pattern is critical; an improper choice results in unsatisfactory performance. This paper proposes a new method of array synthesis that allows the design of a robust broadband beamformer with tunable tradeoff between frequency-invariance and directivity, without the need for imposing a desired beam pattern. The latter is defined as a set of variables that do not depend on frequency and are included in the vector of variables to be optimized. To this end, a suitable cost function has been devised whose minimum can be found in closed form. Therefore, the method is analytical and computationally inexpensive. In addition, a technique that allows obtaining a beam pattern with a linear phase over frequency is described. The results show the effectiveness of the proposed method in designing robust superdirective beam patterns for linear arrays receiving far-field signals, with special attention to microphone arrays of limited aperture.

A closed form solution to the microphone position self-calibration problem

This paper presents a novel algorithm for the automatic 3D localization of a set of microphones in an unknown environment. Given the times of arrival at each microphone of a set of sound events, the approach simultaneously estimates the 3D positions of the sensors and the sources that have generated the events. The only assumption made is that the emission time of the sound events must be known in order to measure the time of flight for each event. A closed form solution is also proposed whenever a sound event coincides with a microphone position. Simulated and real experiments show the validity of the approach for different setups of sensors and number of events.

A stochastic approach to the synthesis of a robust frequency-invariant filter-and-sum beamformer

Frequency-invariant beam patterns (FIBPs) are often required by systems using an array of sensors to process broadband signals. Although several methods have been proposed to design a broadband beamformer (typically characterized by a finite impulse response (FIR) filter for each sensor) with an FIBP, until now, the case in which the spatial aperture is shorter than the involved wavelengths has very rarely been considered. In such a case, the use of a superdirective beam pattern is essential to attaining an efficient system. In this context, robustness to array imperfections and random errors is a very crucial feature. In this paper, a method to design a robust broadband beamformer that produces an FIBP for a data-independent superdirective array is proposed and compared with other potential approaches. The method generates a far-field beam pattern that reproduces the desired profile over a very wide frequency band (also if the array is shorter than the wavelength) and is based on a stochastic approach to the direct synthesis of the FIR filters. The very simple implementation and the resulting robustness of the attained filter-and-sum beamformer to array imperfections increase the applicability of the system. This fact is particularly important in the context of a audio signal processing carried out by microphone arrays, which is the main application considered in this paper.

Audio Surveillance: A Systematic Review

Despite surveillance systems becoming increasingly ubiquitous in our living environment, automated surveillance, currently based on video sensory modality and machine intelligence, lacks most of the time the robustness and reliability required in several real applications. To tackle this issue, audio sensory devices have been incorporated, both alone or in combination with video, giving birth in the past decade, to a considerable amount of research. In this article, audio-based automated surveillance methods are organized into a comprehensive survey: A general taxonomy, inspired by the more widespread video surveillance field, is proposed to systematically describe the methods covering background subtraction, event classification, object tracking, and situation analysis. For each of these tasks, all the significant works are reviewed, detailing their pros and cons and the context for which they have been proposed. Moreover, a specific section is devoted to audio features, discussing their expressiveness and their employment in the above-described tasks. Differing from other surveys on audio processing and analysis, the present one is specifically targeted to automated surveillance, highlighting the target applications of each described method and providing the reader with a systematic and schematic view useful for retrieving the most suited algorithms for each specific requirement.

Heterogeneous Auto-similarities of Characteristics (HASC): Exploiting Relational Information for Classification

Capturing the essential characteristics of visual objects by considering how their features are inter-related is a recent philosophy of object classification. In this paper, we embed this principle in a novel image descriptor, dubbed Heterogeneous Auto-Similarities of Characteristics (HASC). HASC is applied to heterogeneous dense features maps, encoding linear relations by co variances and nonlinear associations through information-theoretic measures such as mutual information and entropy. In this way, highly complex structural information can be expressed in a compact, scale invariant and robust manner. The effectiveness of HASC is tested on many diverse detection and classification scenarios, considering objects, textures and pedestrians, on widely known benchmarks (Caltech-101, Brodatz, Daimler Multi-Cue). In all the cases, the results obtained with standard classifiers demonstrate the superiority of HASC with respect to the most adopted local feature descriptors nowadays, such as SIFT, HOG, LBP and feature co variances. In addition, HASC sets the state-of-the-art on the Brodatz texture dataset and the Daimler Multi-Cue pedestrian dataset, without exploiting ad-hoc sophisticated classifiers.

The synthesis of robust broadband beamformers for equally-spaced linear arrays.

Broadband beamforming applied to superdirective arrays is known to be highly sensitive to transducers characteristics errors. Recently, an effective method to synthesize a robust, broadband, data-independent, filter-and-sum beamformer, which considers the probability distributions of errors, has been proposed. It considers an array with arbitrary lay-out and provides a spatial directivity pattern close to the desired one. Unfortunately, the evaluation of the cost function requires a computational load and an amount of memory that increases very quickly with the transducers and the filter coefficients. As the synthesis process requires an iterative minimization, it may become unacceptably long. In this paper, an alternative procedure for evaluating the same cost function is presented, drastically reducing such problems and without introducing any approximation. The only additional constraint is on the antenna which must be an equally-spaced linear array. This procedure makes it much easier to apply the above-mentioned synthesis method to a wide panorama of practical situations.

Combining multi-pulse excitation and chirp coding in contrast-enhanced ultrasound imaging

The development of techniques to separate the response of the contrast agent from that of the biological tissues is of great importance in ultrasound medical imaging. In the literature, one can find various solutions involving the use of multiple transmitted signals and the weighted sum of related echoes. In this paper, the combination of one of these multi-pulse techniques with a coded excitation is proposed and assessed. Coded excitation has been used mainly to increase the signal-to-noise ratio (SNR) and the penetration depth, provided that a matched filtering is applied in the reception chain. However, it has been shown that a signal with a long duration time also increases the backscattered echoes produced by the microbubbles and, consequently, the contrast-to-tissue ratio. Therefore, the implementation of a multi-pulse technique using a long coded pulse can yield a better contrast-to-tissue ratio and SNR. This paper investigates the combination of the linear chirp pulse with a multi-pulse technique based on the transmission of three pulses. The performance was evaluated using both simulated and real signals, assessing the improvement in the contrast-to-tissue ratio and SNR, the visual quality of the images obtained and the axial accuracy. A comparison with the same multi-pulse technique implemented using a traditional amplitude-modulated pulse revealed that the deployment of a chirp pulse produces several noticeable advantages and only a minor drawback.

Look at Who’s Talking: Voice Activity Detection by Automated Gesture Analysis

This paper proposes an approach for Voice Activity Detection (VAD) based on the automatic measurement of gesturing. The main motivation of the work is that gestures have been shown to be tightly correlated with speech, hence they can be considered a reliable evidence that a person is talking. The use of gestures rather than speech for performing VAD can be helpful in many situation (e.g., surveillance and monitoring in public spaces) where speech cannot be obtained for technical, legal or ethical issues. The results show that the gesturing measurement approach proposed in this work achieves, on a frame-by-frame basis, an accuracy of 71 percent in distinguishing between speech and non-speech.

A Computationally Efficient Procedure for the Design of Robust Broadband Beamformers

Broadband beamformers using superdirective arrays are sensitive to errors in the sensor characteristics. Recently, a method that considers the probability distributions of errors to synthesize a robust optimal beamformer has been proposed. Unfortunately, the entire synthesis process may become unacceptably long due to the computational burden of the cost function. Here, we present an alternative procedure for computing the same cost function, drastically reducing such a problem.