There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

CA : A Cancer Journal for Clinicians

Introduction to linear and nonlinear programming

Bilateral teleoperation: An historical survey

Detection of bone surfaces in ultrasound images would be useful for ultrasound guided orthopedic surgery, biopsy and brachytherapy. However, bones are often poorly visualized with conventional B-mode ultrasound due to speckle, shadowing, reverberation and other artifacts in tissue. In this paper, we investigate two new techniques for the enhancement of bone surface visualization using ultrasound radio frequency (RF) signals, instead of using conventional B-mode images. The first approach uses strain imaging or elastography, and the second method directly monitors the reflected power of the RF signal. The potential of the proposed methods is demonstrated through phantom and in vivo experiments. Experimental results show that the two methods produce satisfactory contrast between bone surfaces and soft tissue, and are suitable for real-time applications. The good performance of these approaches suggests that they have promise in a clinical setting.

P6D-5 Enhancement of Bone Surface Visualization Using Ultrasound Radio-Frequency Signals

To meet the high performance demands of modern robot applications, design variables such as materials, geometry, actuators and sensors must be chosen for optimum performance. This thesis presents a new way of choosing design variables to tune the capabilities of a robot to the needs of an application. The associated proposals are demonstrated through the design of a haptic interface. 
It is argued that isotropy over a given workspace is a good measure of design quality for many high performance applications. A new measure of isotropy, the Global Isotropy Index or GII, is presented which is computed from the singular values of a design matrix. To ensure that the singular values are meaningful, a technique is presented which normalizes and scales the design matrix. The technique removes all physical units from the design matrix and scales it to accommodate an application-dependent performance specification and non-homogeneous actuator capabilities. An algorithm has also been developed that solves for the design parameters that result in the optimum GII. It is so efficient that it can be used to compare the relative isotropies of different robot configurations. 
Performance specifications for a haptic interface are taken from the literature and are augmented by two biomechanical studies. The values obtained are used by the proposed design procedure to select the best of three robots to be used as a haptic pen. The preferred candidate is a novel hybrid design that uses two 3-DOF pantographs to position the ends of a pen shaped end-effector. A prototype with passive roll about the pen axis is built and controlled to simulate three virtual environments including a virtual pencil, a virtual scalpel and a virtual excavator. Its performance characteristics are measured and are used to draw conclusions about the effectiveness of the design procedure. 
Finally, a further performance improvement is sought via redundant actuation. It is shown that the motion range and force capabilities of a coarse-stage robot can be combined with the precision and high-acceleration of a fine-stage robot by connecting them in series and joining their end-effectors by a flexible coupling. A coarse-fine system such as this is expected to narrow the gap between achievable and ideal haptic interface performance.

/pdf/robot-design-optimization-with-haptic-interface-applications-28taet0w28.pdf

Robot design optimization with haptic interface applications

Intraoperative transrectal ultrasound (TRUS) has the potential to improve the outcomes of robotic-assisted laparoscopic radical prostatectomy (RALRP), a common treatment for prostate cancer. Intraoperative TRUS is currently difficult to implement in RALRP because a dedicated assistant with access to the patient is required. Previously described robotic TRUS probe manipulators overcome this limitation but require the surgeon to frequently interrupt their operation to reposition the TRUS probe under remote control. A new method for automatically tracking existing da Vinci surgical instruments with a TRUS imaging plane is presented, based on a rigid registration of the kinematic frames of the da Vinci system and the robotic TRUS probe manipulator. This registration is accomplished by identifying fiducial points on a tissue surface using the tip of a da Vinci manipulator, and does not require any external tracking system. Phantom validation shows an average registration error of 0.95 ± 0.38 mm based on manually locating the manipulator tips in TRUS images. This corresponds to an average tracking error of approximately 1 degree when rotating a parasagittal TRUS imaging plane to track a da Vinci manipulator.

Instrument-based calibration and remote control of intraoperative ultrasound for robot-assisted surgery

A system and method includes applying, by a vibration source, periodic excitation waves to the volume of tissue, the periodic excitation wave having a plurality of phases, acquiring by the CT scanner a first plurality of CT projections at a first phase, the first plurality of CT projections comprising a first CT projection set, acquiring by the CT scanner a second plurality of CT projections at a second phase, wherein the second phase is different than the first phase a second set of CT projections, the second plurality of CT projections comprising a second CT projection set, and determining, based on the first set of CT projections and the second set of CT projections, at least one of a tissue deformation field and a tissue mechanical property of the volume of tissue.

Dynamic computed tomography imaging of elasticity

Surgical instrument segmentation for robot-assisted surgery is needed for accurate instrument tracking and augmented reality overlays. Therefore, the topic has been the subject of a number of recent papers in the CAI community. Deep learning-based methods have shown state-of-the-art performance for surgical instrument segmentation, but their results depend on labelled data. However, labelled surgical data is of limited availability and is a bottleneck in surgical translation of these methods. In this paper, we demonstrate the limited generalizability of these methods on different datasets, including robot-assisted surgeries on human subjects. We then propose a novel joint generation and segmentation strategy to learn a segmentation model with better generalization capability to domains that have no labelled data. The method leverages the availability of labelled data in a different domain. The generator does the domain translation from the labelled domain to the unlabelled domain and simultaneously, the segmentation model learns using the generated data while regularizing the generative model. We compared our method with state-of-the-art methods and showed its generalizability on publicly available datasets and on our own recorded video frames from robot-assisted prostatectomies. Our method shows consistently high mean Dice scores on both labelled and unlabelled domains when data is available only for one of the domains.

Septimiu E. Salcudean

Papers

P6D-5 Enhancement of Bone Surface Visualization Using Ultrasound Radio-Frequency Signals

Robot design optimization with haptic interface applications

Instrument-based calibration and remote control of intraoperative ultrasound for robot-assisted surgery

Dynamic computed tomography imaging of elasticity

Co-generation and Segmentation for Generalized Surgical Instrument Segmentation on Unlabelled Data