Statistical methods for biomedical signal analysis and processing

Statistical modelling and statistical learning theory are two powerful analytical frameworks for analyzing signals and developing efficient processing and classification algorithms. In this thesis, these frameworks are applied for modelling and processing biomedical signals in two different contexts: ultrasound medical imaging systems and primate neural activity analysis and modelling. In the context of ultrasound medical imaging, two main applications are explored: deconvolution of signals measured from a ultrasonic transducer and automatic image segmentation and classification of prostate ultrasound scans. In the former application a stochastic model of the radio frequency signal measured from a ultrasonic transducer is derived. This model is then employed for developing in a statistical framework a regularized deconvolution procedure, for enhancing signal resolution. In the latter application, different statistical models are used to characterize images of prostate tissues, extracting different features. These features are then uses to segment the images in region of interests by means of an automatic procedure based on a statistical model of the extracted features. Finally, machine learning techniques are used for automatic classification of the different region of interests. In the context of neural activity signals, an example of bio-inspired dynamical network was developed to help in studies of motor-related processes in the brain of primate monkeys. The presented model aims to mimic the abstract functionality of a cell population in 7a parietal region of primate monkeys, during the execution of learned behavioural tasks.

Efficient ultrasonic signal processing techniques for aided medical diagnostics

Ultrasound imaging is widely used in medical diagnostics as it is the fastest, least invasive, and least expensive imaging modality. However, ultrasound images are intrinsically difficult to be interpreted. In this scenario, Computer Aided Detection (CAD) systems can be used to support physicians during diagnosis providing them a second opinion. This thesis discusses efficient ultrasound processing techniques for computer aided medical diagnostics, focusing on two major topics: (i) Ultrasound Tissue Characterization (UTC), aimed at characterizing and differentiating between healthy and diseased tissue; (ii) Ultrasound Image Segmentation (UIS), aimed at detecting the boundaries of anatomical structures to automatically measure organ dimensions and compute clinically relevant functional indices. Research on UTC produced a CAD tool for Prostate Cancer detection to improve the biopsy protocol. In particular, this thesis contributes with: (i) the development of a robust classification system; (ii) the exploitation of parallel computing on GPU for real-time performance; (iii) the introduction of both an innovative Semi-Supervised Learning algorithm and a novel supervised/semi-supervised learning scheme for CAD system training that improve system performance reducing data collection effort and avoiding collected data wasting. The tool provides physicians a risk map highlighting suspect tissue areas, allowing them to perform a lesion-directed biopsy. Clinical validation demonstrated the system validity as a diagnostic support tool and its effectiveness at reducing the number of biopsy cores requested for an accurate diagnosis. For UIS the research developed a heart disease diagnostic tool based on Real-Time 3D Echocardiography. Thesis contributions to this application are: (i) the development of an automated GPU based level-set segmentation framework for 3D images; (ii) the application of this framework to the myocardium segmentation. Experimental results showed the high efficiency and flexibility of the proposed framework. Its effectiveness as a tool for quantitative analysis of 3D cardiac morphology and function was demonstrated through clinical validation.

Embedded systems and advanced signal processing for Acousto- Ultrasonic Inspections

Non Destructive Testing (NDT) and Structural Health Monitoring (SHM) are becoming essential in many application contexts, e.g. civil, industrial, aerospace etc., to reduce structures maintenance costs and improve safety. Conventional inspection methods typically exploit bulky and expensive instruments and rely on highly demanding signal processing techniques. The pressing need to overcome these limitations is the common thread that guided the work presented in this Thesis. In the first part, a scalable, low-cost and multi-sensors smart sensor network is introduced. The capability of this technology to carry out accurate modal analysis on structures undergoing flexural vibrations has been validated by means of two experimental campaigns. Then, the suitability of low-cost piezoelectric disks in modal analysis has been demonstrated. To enable the use of this kind of sensing technology in such non conventional applications, ad hoc data merging algorithms have been developed. In the second part, instead, imaging algorithms for Lamb waves inspection (namely DMAS and DS-DMAS) have been implemented and validated. Results show that DMAS outperforms the canonical Delay and Sum (DAS) approach in terms of image resolution and contrast. Similarly, DS-DMAS can achieve better results than both DMAS and DAS by suppressing artefacts and noise. To exploit the full potential of these procedures, accurate group velocity estimations are required. Thus, novel wavefield analysis tools that can address the estimation of the dispersion curves from SLDV acquisitions have been investigated. An image segmentation technique (called DRLSE) was exploited in the k-space to draw out the wavenumber profile. The DRLSE method was compared with compressive sensing methods to extract the group and phase velocity information. The validation, performed on three different carbon fibre plates, showed that the proposed solutions can accurately determine the wavenumber and velocities in polar coordinates at multiple excitation frequencies.

A skin surface characterization system based on capacitive image analysis

During the last few years, several methods have been proposed in order to study and to evaluate characteristic properties of the human skin by using non-invasive approaches. Mostly, these methods cover aspects related to either dermatology, to analyze skin physiology and to evaluate the effectiveness of medical treatments in skin diseases, or dermocosmetics and cosmetic science to evaluate, for example, the effectiveness of anti-aging treatments. To these purposes a routine approach must be followed. Although very accurate and high resolution measurements can be achieved by using conventional methods, such as optical or mechanical profilometry for example, their use is quite limited primarily to the high cost of the instrumentation required, which in turn is usually cumbersome, highlighting some of the limitations for a routine based analysis. This thesis aims to investigate the feasibility of a noninvasive skin characterization system based on the analysis of capacitive images of the skin surface. The system relies on a CMOS portable capacitive device which gives 50 micron/pixel resolution capacitance map of the skin micro-relief. In order to extract characteristic features of the skin topography, image analysis techniques, such as watershed segmentation and wavelet analysis, have been used to detect the main structures of interest: wrinkles and plateau of the typical micro-relief pattern. In order to validate the method, the features extracted from a dataset of skin capacitive images acquired during dermatological examinations of a healthy group of volunteers have been compared with the age of the subjects involved, showing good correlation with the skin ageing effect. Detailed analysis of the output of the capacitive sensor compared with optical profilometry of silicone replica of the same skin area has revealed potentiality and some limitations of this technology. Also, applications to follow-up studies, as needed to objectively evaluate the effectiveness of treatments in a routine manner, are discussed.

Variational and learning models for image and time series inverse problems

Inverse problems are at the core of many challenging applications. Variational and learning models provide estimated solutions of inverse problems as the outcome of specific reconstruction maps. In the variational approach, the result of the reconstruction map is the solution of a regularized minimization problem encoding information on the acquisition process and prior knowledge on the solution. In the learning approach, the reconstruction map is a parametric function whose parameters are identified by solving a minimization problem depending on a large set of data. In this thesis, we go beyond this apparent dichotomy between variational and learning models and we show they can be harmoniously merged in unified hybrid frameworks preserving their main advantages. We develop several highly efficient methods based on both these model-driven and data-driven strategies, for which we provide a detailed convergence analysis. The arising algorithms are applied to solve inverse problems involving images and time series. For each task, we show the proposed schemes improve the performances of many other existing methods in terms of both computational burden and quality of the solution. In the first part, we focus on gradient-based regularized variational models which are shown to be effective for segmentation purposes and thermal and medical image enhancement. We consider gradient sparsity-promoting regularized models for which we develop different strategies to estimate the regularization strength. Furthermore, we introduce a novel gradient-based Plug-and-Play convergent scheme considering a deep learning based denoiser trained on the gradient domain. In the second part, we address the tasks of natural image deblurring, image and video super resolution microscopy and positioning time series prediction, through deep learning based methods. We boost the performances of supervised, such as trained convolutional and recurrent networks, and unsupervised deep learning strategies, such as Deep Image Prior, by penalizing the losses with handcrafted regularization terms.