Catálogo de publicaciones - libros

We introduce a novel framework, called , to store segmentation references for medical images, combine multiple references, and measure the discrepancy between a segmented object and a reference. The core feature is the use of efficiently encoded confidence maps, which reflect the local variations of blur and the presence of nearby objects. Local confidence values are defined from expert user input, and used to define a new discrepancy error measure, aimed to be directly interpreted quantitatively and qualitatively. We illustrate the use of this framework to compare different segmentation methods and tune a method’s parameters.

Dual-energy mammography can suppress the contrast between adipose and glandular tissues and improve the detectability of microcalcifications (MCs). In clinical dual-energy mammography, imaging object is human breast, while in calibration measurements, only phantoms of breast-tissue-equivalent material can be used. The composition and density differences between calibration materials and human breast bring the differences of linear attenuation coefficient which lead to the calculation errors in dual-energy imaging. In this paper, the magnitude of MC thickness error from calibration materials has been analyzed using a first-order propagation of error analysis. This analysis shows that the thickness error from calibration materials ranges from dozens to thousands of microns which can not be ignored when carrying out dual-energy calculations. The evaluation of several popular phantoms shows that it is of great importance to adopt the phantom materials approaching human breast most.

When imaging tendons and cartilage in a MRI scanner, an increase in signal intensity is observed when they are oriented at 55 degrees with respect to (the “magic angle”). There is a clear clinical importance for considering this effect as part of the diagnosis of orthopaedic and other injury. Experimental studies of this phenomenon have been made harder by practical difficulties of tissue positioning and orientation in the confined environment of cylindrical scanners. An MRI compatible mechatronic system has been developed to position a variety of limbs inside the field of view of the scanner, to be used as a diagnostic and research tool. It is actuated with a novel pneumatic motor comprised of a heavily geared down air turbine, and is controlled in a closed loop using standard optical encoders. MR compatibility is demonstrated as well as the results of preliminary trials used to image the Achilles tendon of human volunteers at different orientations. A 4 to 13 fold increase in signal at the tendon is observed at the magic angle.

We present a novel method to track a guidewire in cardiac x-ray video. Using variational calculus, we derive differential equations that deform a spline, subject to intrinsic and extrinsic forces, so that it matches the image data, remains smooth, and preserves an a priori length. We analytically derive these equations from first principles, and show how they include tangential terms, which we include in our model. To address the poor contrast often observed in x-ray video, we propose using phase congruency as an image-based feature. Experimental results demonstrate the success of the method in tracking guidewires in low contrast x-ray video.

Manipulating small objects such as needles, screws or plates inside the human body during minimally invasive surgery can be very difficult for less experienced surgeons, due to the loss of 3D depth perception. This paper presents an approach for tracking a suturing needle using a standard endoscope. The resulting pose information of the needle is then used to generate artificial 3D cues on the 2D screen to optimally support surgeons during tissue suturing. Additionally, if an external tracking device is provided to report the endoscope’s position, the suturing needle can be tracked in a hybrid fashion with sub-millimeter accuracy. Finally, a visual navigation aid can be incorporated, if a 3D surface is intraoperatively reconstructed from video or registered from preoperative imaging.

We study the effect of anisotropic noise on target registration error (TRE) by using a tracked and calibrated stylus tip as the fiducial registration application. We present a simple, efficient unscented Kalman filter algorithm that is suitable for fiducial registration even with a small number of fiducials. We also derive an equation that predicts TRE under anisotropic noise. The predicted TRE values are shown to closely match the simulated TRE values achieved using our UKF-based algorithm.

For decades, conventional 2D-roadmaping has been the method of choice for image-based guidewire navigation during endovascular procedures. Only recently have 3D-roadmapping techniques become available that are based on the acquisition and reconstruction of a 3D image of the vascular tree. In this paper, we present a new image-based navigation technique called RoRo (Rotational Roadmapping) that eliminates the guess-work inherent to the conventional 2D method, but does not require a 3D image. Our preliminary clinical results show that there are situations in which RoRo is preferred over the existing two methods, thus demonstrating potential for filling a clinical niche and complementing the spectrum of available navigation tools.

This paper presents a method for bronchoscope tracking without any fiducial markers using an ultra-tiny electromagnetic tracker (UEMT) for a bronchoscopy guidance system. The proposed method calculates the transformation matrix, which shows the relationship between the coordinates systems of the pre-operative CT images and the UEMT, by registering bronchial branches segmented from CT images and points measured by the UEMT attached at the tip of a bronchoscope. We dynamically compute the transformation matrix for every pre-defined number of measurements. We applied the proposed method to a bronchial phantom in several experimental environments. The experimental results showed the proposed method can track a bronchoscope camera with about 3.3mm of target registration error (TRE) for wood table environment and 4.0mm of TRE for examination table environment.

For current surgical navigation systems optical tracking is state of the art. The accuracy of these tracking systems is currently determined statically for the case of full visibility of all tracking targets. We propose a dynamic determination of the accuracy based on the visibility and geometry of the tracking setup. This real time estimation of accuracy has a multitude of applications. For multiple camera systems it allows reducing line of sight problems and guaranteeing a certain accuracy. The visualization of these accuracies allows surgeons to perform the procedures taking to the tracking accuracy into account. It also allows engineers to design tracking setups interactively guaranteeing a certain accuracy.

Our model is an extension to the state of the art models of Fitzpatrick et al.[1] and Hoff et al. [2]. We model the error in the camera sensor plane. The error is propagated using the internal camera parameter, camera poses, tracking target poses, target geometry and marker visibility, in order to estimate the final accuracy of the tracked instrument.

With the increasing sophistication of surgical robots, the use of motion stabilisation for enhancing the performance of micro-surgical tasks is an actively pursued research topic. The use of mechanical stabilisation devices has certain advantages, in terms of both simplicity and consistency. The technique, however, can complicate the existing surgical workflow and interfere with an already crowded MIS operated cavity. With the advent of reliable vision-based real-time and techniques on 3D-deformation recovery, current effort is being directed towards the use of optical based techniques for achieving adaptive motion stabilisation. The purpose of this paper is to assess the effect of virtual stabilization on foveal/parafoveal vision during robotic assisted MIS. Detailed psychovisual experiments have been performed. Results show that stabilisation of the whole visual field is not necessary and it is sufficient to perform accurate motion tracking and deformation compensation within a relatively small area that is directly under foveal vision. The results have also confirmed that under the current motion stabilisation regime, the deformation of the periphery does not affect the visual acuity and there is no indication of the deformation velocity of the periphery affecting foveal sensitivity. These findings are expected to have a direct implication on the future design of visual stabilisation methods for robotic assisted MIS.