Fereshteh Aalamifar

Real-time tracking of a bevel-tip needle with varying insertion depth: Toward teleoperated MRI-guided needle steering

This study presents one of the enabling technologies for teleoperated bevel-tip needle steering under real-time MRI guidance i.e. capability of tracking the needle with higher accuracy and bandwidth than real-time MRI. Three fibers, each with three Fiber Bragg Gratings (FBG) were embedded into a 0.6 mm inner stylet of a 20G MRI-compatible biopsy needle. The axial force caused by the bevel-tip was considered in the analysis using beam-column theory. Since the insertion depth is varying, the minimum number of sensors and their optimal locations in the fibers were determined such that the tip position error estimation is below 0.5 mm for all insertion depths. A practical and accurate calibration method for the apparatus is presented. The instrumented needle was fabricated to fit in the needle driver unit of a MRI-compatible needle steering robot. The tracking apparatus was calibrated, including compensation for temperature changes in tissue during insertion. Experimental results showed needle tip tracking error below 0.5 mm at different insertion depths. Real-time 3D shape of the needle was visualized in 3D Slicer yielding navigation of the needle in real-time.

Multiple cognitive nodes in wireless sensor networks

Recently, cognitive wireless sensor networks (CWSN) have been proposed to enhance the overall functionality of WSNs based on application needs. Cognition refers to the process of making decisions and acting based on the network conditions in order to achieve end-to-end goals of the network. In the previously proposed architecture, a new node - Cognitive Node (CN)-, which is responsible for managing the network, was introduced to the network. Since WSNs normally consist of a huge number of Sensor Nodes (SNs), there is a high demand to have more than one CN. In this paper, challenges of adding multiple CNs are investigated and a new architecture is proposed. Then, the proposed architecture is evaluated by implementing it with CC2430 boards (Texas Instrument boards). The results of our experiments showed that the proposed architecture can function more efficiently by increasing the number of SNs in each subclass.

Toward teleoperated needle steering under continuous MRI guidance for prostate percutaneous interventions.

To propose a human‐operated in‐room master–slave bevel‐tip needle steering system under continuous MRI guidance for prostate biopsy, in which the patient is kept in the scanner at all times and the process of needle placement is under continuous control of the physician.

Enabling technologies for robot assisted ultrasound tomography: system setup and calibration

In this study, we are proposing a robot-assisted ultrasound tomography system that can offer soft tissue tomographic imaging and deeper or faster scan of the anatomy. This system consists of a robot-held ultrasound probe that tracks the position of another freehand probe, trying to align with it. One of the major challenges is achieving proper alignment of the two ultrasound probes. To enable proper alignment, two ultrasound calibrations and one hand-eye calibration are required. However, the system functionality and design is such that the ultrasound calibrations have become a challenge. In this paper, after providing an overview of the proposed robotic ultrasound tomography system, we focus on the calibrations problem. The results of the calibrations show a point reconstruction precision of a few millimeters for the current prototype, and the two images have at least 50% overlap visually; confirming the feasibility of such a system relying on accurate probe alignments.

Robot-assisted mirror ultrasound scanning for deep venous thrombosis detection using RGB-D sensor

Deep Venous Thrombosis (DVT) is a major cause of morbidity and mortality. Manually scanning with Ultrasound (US) probe brings heavy work load for the sonographers. This paper proposes a novel “Mirror robotic US scanning system. The system is composed of two-arm robot, linear US probes for master and slave side, Kinect sensor as a vision servo. On the master side, sonographers hold one robot arm and operate probe to inspect one leg. On the slave side, the robot follows the master probe on the navigation of Kinect and scans the other leg. 3D images of legs are segmented and register to get mirror matrix. Both Clustered Viewpoint Feature Histogram (CVFH) descriptors of segmented probe and CAD (Computer-Aided Design) training data were calculated to probe recognition. The leg phantom platform was built up. The mirror matrixes were obtained. The correlation coefficients between the two legs are calculated. Times of ICP (Iterative Closest Point) have been calculated for the platform. Results from the initial experiment indicate the idea is feasible and promising greatly by improving the inspection efficiently. Clinically, the method can be implemented for pre-operative procedures to predict the risk of DVT, and this may improve the US scanning efficacy.

Co-robotic ultrasound tomography: dual arm setup and error analysis

Ultrasound (US) tomography enables quantitative measurement of acoustic properties. Robot assisted ultrasound tomography system enables alignment of two US probes. The alignment is done automatically by the robotic arm so that tomographic reconstruction of more anatomies becomes possible. In this study, we propose a new system setup for robot assistance in US tomographic imaging. This setup includes two robotic arms holding two US probes. One of the robotic arms is operated by the sonographer to determine the desired location for the tomographic imaging; this probe can also provide the B-mode US image during the search. The other robotic arm can then move automatically to align the two probes. One of the probes will act as transmitter and the other one as receiver to enable tomographic imaging. We provide an overview of the system setup and components together with the calibration procedures. In an attempt to provide a complete framework for the tomography system, we also provide a sample tomographic reconstruction method that can reconstruct speed of sound image using two aligned linear US probes. The reconstruction algorithm is however very prone to alignment inaccuracies. We provide an error propagation analysis to provide an estimation of the overall alignment error and then show the effect of the in-plane translational error in the tomographic reconstruction.

An iterative algorithmic implementation of input-output finite state approximations

The problem of finding finite state models of systems with quantized inputs and outputs has received much deserved attention. In particular, a notion of ρ/μ approximation was proposed and shown to be relevant to the problem of control synthesis. In this paper, we revisit a recently developed constructive algorithm for generating ρ/μ approximations for a class of systems, and we propose and analyze several algorithms for improving the computational efficiency and memory requirements of the construction. We demonstrate the use of this approach for synthesizing certified-by-design controllers for a simple illustrative example with reachability type specifications.

Enabling technologies for robot assisted ultrasound tomography

Currently available ultrasound (US) tomography systems suggest utilizing cylindrical transducers that can be used for a specific organ. In this paper, our focus is on an alternative way of creating US tomographic images that could be used for other anatomies and more general applications. This system consists of two conventional US probes facing each other while one or several of the transducers in one probe can act as the transmitter and the rest as the receiver. Aligning the two US probes is a challenging task. To address this issue, we propose a robot assisted US tomography system in which one probe is operated freehanded and another by a robotic arm. In this paper, enabling technologies for this system are described. With the current prototype, a reconstruction precision of 4.12, 1.73, and 2.23 mm for the three calibrations, and an overall alignment repeatability in the range of 5–9 mm were achieved. Copyright

In vivo visualization of robotically implemented synthetic tracked aperture ultrasound (STRATUS) imaging system using curvilinear array

Synthetic aperture for ultrasound is a technique utilizing a wide aperture in both transmit and receive to enhance the ultrasound image quality. The limitation of synthetic aperture is the maximum available aperture size limit determined by the physical size of ultrasound probe. We propose Synthetic-Tracked Aperture Ultrasound (STRATUS) imaging system to overcome the limitation by extending the beamforming aperture size through ultrasound probe tracking. With a setup involving a robotic arm, the ultrasound probe is moved using the robotic arm, while the positions on a scanning trajectory are tracked in real-time. Data from each pose are synthesized to construct a high resolution image. In previous studies, we have demonstrated the feasibility through phantom experiments. However, various additional factors such as real-time data collection or motion artifacts should be taken into account when the in vivo target becomes the subject. In this work, we build a robot-based STRATUS imaging system with continuous data collection capability considering the practical implementation. A curvilinear array is used instead of a linear array to benefit from its wider capture angle. We scanned human forearms under two scenarios: one submerged the arm in the water tank under 10 cm depth, and the other directly scanned the arm from the surface. The image contrast improved 5.51 dB, and 9.96 dB for the underwater scan and the direct scan, respectively. The result indicates the practical feasibility of STRATUS imaging system, and the technique can be potentially applied to the wide range of human body.

Classification of kidney and liver tissue using ultrasound backscatter data

Ultrasound (US) tissue characterization provides valuable information for the initialization of automatic segmentation algorithms, and can further provide complementary information for diagnosis of pathologies. US tissue characterization is challenging due to the presence of various types of image artifacts and dependence on the sonographer’s skills. One way of overcoming this challenge is by characterizing images based on the distribution of the backscatter data derived from the interaction between US waves and tissue. The goal of this work is to classify liver versus kidney tissue in 3D volumetric US data using the distribution of backscatter US data recovered from end-user displayed Bmode image available in clinical systems. To this end, we first propose the computation of a large set of features based on the homodyned-K distribution of the speckle as well as the correlation coefficients between small patches in 3D images. We then utilize the random forests framework to select the most important features for classification. Experiments on in-vivo 3D US data from nine pediatric patients with hydronephrosis showed an average accuracy of 94% for the classification of liver and kidney tissues showing a good potential of this work to assist in the classification and segmentation of abdominal soft tissue.