Renato Seiji Tavares

Simulated annealing with adaptive neighborhood: A case study in off-line robot path planning

Simulated annealing (SA) is an optimization technique that can process cost functions with degrees of nonlinearities, discontinuities and stochasticity. It can process arbitrary boundary conditions and constraints imposed on these cost functions. The SA technique is applied to the problem of robot path planning. Three situations are considered here: the path is represented as a polyline; as a Bezier curve; and as a spline interpolated curve. In the proposed SA algorithm, the sensitivity of each continuous parameter is evaluated at each iteration increasing the number of accepted solutions. The sensitivity of each parameter is associated to its probability distribution in the definition of the next candidate.

Registration of temporal sequences of coronal and sagittal MR images through respiratory patterns

Abstract This work discusses the determination of the breathing patterns in time sequence of images obtained from magnetic resonance (MR) and their use in the temporal registration of coronal and sagittal images. The registration is made without the use of any triggering information and any special gas to enhance the contrast. The temporal sequences of images are acquired in free breathing. The real movement of the lung has never been seen directly, as it is totally dependent on its surrounding muscles and collapses without them. The visualization of the lung in motion is an actual topic of research in medicine. The lung movement is not periodic and it is susceptible to variations in the degree of respiration. Compared to computerized tomography (CT), MR imaging involves longer acquisition times and it is preferable because it does not involve radiation. As coronal and sagittal sequences of images are orthogonal to each other, their intersection corresponds to a segment in the three-dimensional space. The registration is based on the analysis of this intersection segment. A time sequence of this intersection segment can be stacked, defining a two-dimension spatio-temporal (2DST) image. The algorithm proposed in this work can detect asynchronous movements of the internal lung structures and lung surrounding organs. It is assumed that the diaphragmatic movement is the principal movement and all the lung structures move almost synchronously. The synchronization is performed through a pattern named respiratory function. This pattern is obtained by processing a 2DST image. An interval Hough transform algorithm searches for synchronized movements with the respiratory function. A greedy active contour algorithm adjusts small discrepancies originated by asynchronous movements in the respiratory patterns. The output is a set of respiratory patterns. Finally, the composition of coronal and sagittal image pairs that are in the same breathing phase is realized by comparing of respiratory patterns originated from diaphragmatic and upper boundary surfaces. When available, the respiratory patterns associated to lung internal structures are also used. The results of the proposed method are compared with the pixel-by-pixel comparison method. The proposed method increases the number of registered pairs representing composed images and allows an easy check of the breathing phase.

Lung Movement Determination in Temporal Sequences of MR Images Using Hough Transform and Interval Arithmetics

Abstract Since the lung is an organ whose real movement has never been seen, as it is totally dependent on the rib cage and collapses without them, the objective of the present work is to develop a method to study its movement. The visualization of the lung in motion is an actual topic of research in medicine. Computerized Tomography (CT) can obtain spatio–temporal images of the heart by synchronizing with electrocardiographic waves. The field of view (FOV) of the heart is small when compared to the lungs FOV. The lungs movement is not periodic and it is susceptible to variations in the degree of respiration. Compared to CT, Magnetic Resonance (MR) imaging involves longer acquisition times and it is preferable because they do not involve radiation. The breathing is associated to a standard respiratory function, and through 2D image processing, edge detection, Hough transform and interval arithmetics, respiratory functions are obtained and, consequently, the position of the points in time are estimated. A standard respiratory function must be provided as input to the modified Hough transform. The Hough transform uses a discrete quantization space, implemented as an accumulator matrix. An improved Hough transform is proposed by using interval arithmetics in the accumulation phase of the algorithm. The discreteness of the quantization space is considered in the algorithm, and its consequences are minimized. This work showed more consistent results when compared to previous approaches. Copyright ©2009 IFAC.

Discretization Error and the EIT Forward Problem

Abstract Electrical impedance tomography is a portable imaging technique in which the image represents internal conductivities within a body. Electrical measurements are made at the body surface, and the internal conductivities are calculated. It is an inverse problem that can be solved by comparing simulated results obtained from numerical simulations performed by the forward problem and measured data. In this approach, the forward problem has a very important role. The forward problem can be solved by the finite element method, and it is mainly influenced by the mesh creation algorithm and the electrode model. This work proposes a mesh that has more elements in the boundary and fewer elements in the center. The electrode model is approximately a rectangular element in which the potentials at the external nodes are considered the same. This fact can reduce the number of variables as it will be shown. The proposed mesh creation algorithm is analyzed according to the discretization error theory. It is concluded that meshes with higher density in the external ring have smaller discretization errors.

Temporal segmentation of lung region from MRI sequences using multiple active contours

Segmentation of the lung is particularly difficult because of the large variation in image quality. A modified Hough transform in combination with a mask creation algorithm can robustly determine synchronous respiratory patterns. The synchronicity restriction is relaxed by applying a greedy active contour algorithm. The respiratory patterns define a point cloud near the lung region boundary representing a subjective contour. The gravitation vector field (GVF) active contour algorithm is used to create an initial segmentation exclusively based on the point cloud. A final active contours algorithm is executed to adjust the boundary to the images. The algorithm was tested with healthy subjects and COPD patients, and the result was checked through temporal registration of coronal and sagittal images.

Temporal segmentation of lung region MR image sequences using hough transform

In this work, segmentation is an intermediate step in the registration and 3D reconstruction of the lung, where the diaphragmatic surface is automatically and robustly isolated. Usually, segmentation methods are interactive and use different strategies to combine the expertise of humans and computers. Segmentation of lung MR images is particularly difficult because of the large variation in image quality. The breathing is associated to a standard respiratory function, and through 2D image processing, edge detection and Hough transform, respiratory patterns are obtained and, consequently, the position of points in time are estimated. Temporal sequences of MR images are segmented by considering the coherence in time. This way, the lung silhouette can be determined in every frame, even on frames with obscure edges. The lung region is segmented in two steps: a mask containing the lung region is created, and the Hough transform is applied exclusively to mask pixels. The shape of the mask can have a large variation, and the modified Hough transform can handle such shape variation. The result was checked through temporal registration of coronal and sagittal images.

COMPARISON OF DIFFERENT TECHNIQUES FOR ELECTRICAL IMPEDANCE TOMOGRAPHY RECONSTRUCTION USING SIMULATED ANNEALING AND GPU PARALLELIZATION

Electrical Impedance Tomography (EIT) is an imaging technique that attempts to reconstruct the conductivity distribution inside an object from electrical currents and potentials applied and measured at its surface. The EIT reconstruction problem is approached as an optimization problem. This optimization problem can be solved using Simulated Annealing (SA), but at a high computational cost. To reduce the computational load, it is possible to use an incomplete evaluation of the objective function. Two objective functions are analyzed and compared: Euclidian distance and least square minimization. The Euclidian distance minimization showed to present an outside-in behavior, determining the impedance of the external elements first, similar to a layer striping algorithm. It also presents the impact of using GPU for parallelizing matrixvector multiplication. Results with experimental data are presented.

Detecting function patterns with interval Hough transform

The Hough transform is a feature extraction technique used in image analysis, computer vision and digital image processing. Usually it is used for detecting straight lines and curves. The purpose of the technique is to find imperfect instances of patterns within a certain class of shapes by a voting procedure. An improved Hough transform is proposed by using interval arithmetics in the accumulation phase of the algorithm. The discreteness of the quantization space is considered. The use of interval arithmetics showed to be simpler than other statistical proposals. It is presented a modified Hough transform for detecting function patterns implemented with interval arithmetics. This modified Hough transform was used to analyze the movement of the lung in MR temporal sequence of images. The results obtained with the interval arithmetics implementation were much better when compared with the conventional Hough transform implementation.