We introduce a wearable-based system for real-time ECG anomaly detection and contextual interpretation within a mobile-health framework. Twenty-four-hour Holter ECG data are synchronized over wireless/mobile networks with e.g. Apple Health streams (iPhone + iWatch), including activity states (walking, running, resting, sleeping) and heart rate history. A hybrid preprocessing pipeline extracts instantaneous frequency (Hilbert), spectral entropy, and RMS energy, concatenated into fixed-length multichannel tensors for deep-learning models deployed via edge or cloud SaaS. The model detects critical cardiac anomalies correlating each with user activity and exertion context. This multimodal approach distinguishes physiological deviations during motion from pathological events at rest or sleep and suppresses motion artifacts. Experiments with subjects wearing both Holter and Apple devices demonstrate improved sensitivity and specificity versus ECG-only baselines. Our system exemplifies wearable computing, mobile health, ML-enabled mobile systems, and edge/cloud mobile analytics. Fig. 1 shows a complete system for recording and classifying ECG signals, including a Holter ECG with electrodes, a smartphone and a smartwatch [1].

The main goal of the research was to develop an effective, highresolution tomographic apparatus capable of non-invasively capturing real-time internal images of industrial tank reactors. For this purpose, a prototype of an ultrasonic tomograph (UST) was developed, which combines innovative design solutions and modern algorithmic techniques. A special feature of the presented solution is the use of a neural network with an unusual architecture. A deep, multi-branch neural network consisting of two inputs was used. The first input is a 120-element vector (sequence) of raw measurements. The third input consists of three sequences obtained as a result of the transformation of raw measurements: instantenous frequency (IF), approximation coefficients (Ca), and detail coefficients (Cd). The prototype was tested on a real model. The tomographic reconstructions obtained using the innovative neural architecture were compared with images obtained using a standard neural network. The results clearly confirm the high effectiveness of the presented approach.

Ultrasound computer tomography (USCT) represents a medical imaging modality designed to visualize alterations in the speed of ultrasonic waves. The primary objective of the study presented was to devise a lightweight, portable, and cost-effective tomographic device capable of non-invasively capturing internal images of the human brain in real-time. To achieve this aim, a prototype ultrasonic tomograph was developed, comprising a lightweight head hoop integrated with ultrasonic transducers and a tomograph unit. Ultrasonic measurements were transformed into images using a heterogeneous convolutional neural network (CNN). The USCT system was engineered to facilitate wireless communication between the sensors embedded within the wearable head cap and the tomographic apparatus.

The primary objective of the research is the development of a lightweight and cost-effective headband-style tomographic apparatus capable of non-invasively capturing real-time internal cerebral images. A prototype of an ultrasonic tomograph was engineered, comprising a lightweight cranial band synergized with ultrasonic transducers and the tomographic system. Ultrasonic measurements were transmuted into visualizations via a heterogeneous convolutional neural network (CNN). The Ultrasonic Computed Tomography (USCT) architecture was conceived to facilitate untethered data interchange between the head-worn sensor array and the tomographic machinery.

Industrial ultrasonic tomography (UT) possesses unique advantages in multiphase medium imaging and has received broad attention. In this article, a novel transmission/reflection dual-mode image reconstruction algorithm based on information fusion is proposed. The transmissive attenuation and reflective time-delay information are both integrated into an improved Lagrange framework with the weighted least square transformation of objective function, which is then solved by a pair of coupled preconditioned gradient approaches. Experiment results show that the proposed algorithm performs better than existing image fusion strategies in terms of accuracy (average relative error 0.456, average correlation coefficient 0.870) and robustness (average standard derivation of relative error and correlation coefficient 0.056 and 0.041). Accordingly, the dual-mode UT approach is proved feasible to provide more accurate image of biphasic medium distribution.

Electrical impedance tomography (EIT) is considered as a potential candidate for brain stroke imaging due to its compactness and potential use in bedside and emergency settings. The electrode-skin contact impedance and low conductivity of skull pose some practical challenges to the EIT head imaging. This paper studies the application of capacitively coupled electrical impedance tomography (CCEIT) in brain imaging for the first time. CCEIT is a new contactless EIT technique which uses voltage excitation without direct contact with the skin, as oppose to directly injecting the current to the skin in EIT. Because the safety issue of a new technique should be strictly treated, simulation work based on a simplified head model was carried out to investigate the safety aspects of CCEIT. By comparing with the standard EIT excited by a typical safe current level used in brain imaging, the safe excitation reference of CCEIT is obtained. This is done by comparing the maximum level of internal electrical field (internal current density) of EIT and that of CCEIT. Simulation results provide useful knowledge of excitation signal level of CCEIT and also show a critical comparison with traditional EIT. Practical experiments were carried out with a 12-electrode CCEIT phantom, saline, and carrot samples. Experimental results show the feasibility and potential of CCEIT for stroke imaging. In this paper, the anomaly diameter resolution is 10 mm (1/18 of the phantom diameter), which indicates that small-volume stroke could be detected. This is achieved by a low excitation voltage of 1 V, showing the possibility of even better performance when higher but yet safe level of excitation voltages is used.

MIT's (magnetic induction tomography) low-cost and noninvasive features can offer great excitement and potential to address many challenging problems that exist in the current industrial/medical applications. The foundation development of MIT has been made in the past ten years. Many more advancements can be expected in the next decade, including the first commercialized MIT system for industrial or medical application. It will certainly contribute some impact to the current imaging technology.

Magnetic induction tomography (MIT) attempts to image the electrical and magnetic characteristics of a target using impedance measurement data from pairs of excitation and detection coils. This inverse eddy current problem is nonlinear and also severely ill posed so regularization is required for a stable solution. A regularized Gauss-Newton algorithm has been implemented as a nonlinear, iterative inverse solver. In this algorithm, one needs to solve the forward problem and recalculate the Jacobian matrix for each iteration. The forward problem has been solved using an edge based finite element method for magnetic vector potential A and electrical scalar potential V, a so called A, A - V formulation. A theoretical study of the general inverse eddy current problem and a derivation, paying special attention to the boundary conditions, of an adjoint field formula for the Jacobian is given. This efficient formula calculates the change in measured induced voltage due to a small perturbation of the conductivity in a region. This has the advantage that it involves only the inner product of the electric fields when two different coils are excited, and these are convenient computationally. This paper also shows that the sensitivity maps change significantly when the conductivity distribution changes, demonstrating the necessity for a nonlinear reconstruction algorithm. The performance of the inverse solver has been examined and results presented from simulated data with added noise.

Image reconstruction in electrical impedance tomography is an ill-posed nonlinear inverse problem. Linearization techniques are widely used and require the repeated solution of a linear forward problem. To account correctly for the presence of electrodes and contact impedances, the so-called complete electrode model is applied. Implementing a standard finite element method for this particular forward problem yields a linear system that is symmetric and positive definite and solvable via the conjugate gradient method. However, preconditioners are essential for efficient convergence. Preconditioners based on incomplete factorization methods are commonly used but their performance depends on user-tuned parameters. To avoid this deficiency, we apply black-box algebraic multigrid, using standard commercial and freely available software. The suggested solution scheme dramatically reduces the time cost of solving the forward problem. Numerical results are presented using an anatomically detailed model of the human head.