We have developed a technique to study how good computers can be at diagnosing gastrointestinal lesions from regular (white light and narrow banded) colonoscopic videos compared two levels of clinical knowledge (expert beginner). Our includes novel tissue classification approach which may save clinician's time by avoiding chromoendoscopy, time-consuming staining procedure using indigo carmine. also discriminates the severity individual in patients with many polyps, so that gastroenterologist...

This paper proposes a grid impedance monitoring system for distributed power generation electronic interfaces. The estimates the equivalent and voltage source from measurements performed at point of common coupling. estimation algorithm is based on recursive least-squares implemented in complex field. Simultaneously, evaluates quality estimation, minimizing its influence detecting islanding situations. proposed performance has been evaluated under experimental testing.

This paper presents a current control scheme for voltage-source converter connected to the grid through an <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$LCL$</tex></formula> filter. The proposed is based on linear quadratic (LQ) Gaussian (LQG) servo controller technique, which combines LQ regulator and Kalman filter (KF). use of KF instead Luenberger estimator improves quality estimation state...

This paper presents a novel approach for indoor acoustic source localization using microphone arrays and based on Convolutional Neural Network (CNN). The proposed solution is, to the best of our knowledge, first published work in which CNN is designed directly estimate three dimensional position an source, raw audio signal as input information avoiding use hand crafted features. Given limited amount available data, we propose this training strategy two steps. We train network semi-synthetic...

We study a problem that we call Shape-from-Template, which is the of reconstructing shape deformable surface from single image and 3D template. Current methods in literature address case isometric deformations, relax isometry constraint to convex inextensibility constraint, solved using so-called maximum depth heuristic. these zeroth-order since they use point locations (the differential structure) solve inference perspective image. propose novel class first-order. The key idea both their...

A major research area in Computer Assisted Intervention (CAI) is to aid laparoscopic surgery teams with Augmented Reality (AR) guidance. This involves registering data from other modalities such as MR and fusing it the video real-time, reveal location of hidden critical structures. We present first system for AR guided uterus. works pre-operative or CT monocular laparoscopes, without requiring any additional interventional hardware optical trackers. novel robust solutions two main...

The massive deployment of smart meters in most Western countries recent decades has allowed the creation and development a significant variety applications, mainly related to efficient energy management. information provided about consumption also been dedicated areas social work health. In this context, are considered single-point non-intrusive sensors that might be used monitor behaviour activity patterns people living household. This describes design short-term behavioural alarm generator...

Measuring the time delay of arrival (TDOA) between a set sensors is basic setup for many applications, such as localization or signal beamforming. This paper presents TDOA matrices, which are built from noise-free measurements, not requiring knowledge sensor array geometry. We prove that matrices rank-two and have special singular value decomposition leads to compact linear parametric representation. Properties applied in this perform denoising, by finding matrix closest composed with noisy...

We study Isometric Non-Rigid Shape-from-Motion (Iso-NRSfM): given multiple intrinsically calibrated monocular images, we want to reconstruct the time-varying 3D shape of a thin-shell object undergoing isometric deformations. show that Iso-NRSfM is solvable from local warps, inter-image geometric transformations. propose new theoretical framework based on Riemmanian manifold represent unknown surfaces as embeddings camera's retinal plane. This allows us use manifold's metric tensor and...

This paper presents a novel approach for indoor acoustic source localization using microphone arrays and based on Convolutional Neural Network (CNN). The proposed solution is, to the best of our knowledge, first published work in which CNN is designed directly estimate three dimensional position an source, raw audio signal as input information avoiding use hand crafted features. Given limited amount available data, we propose this training strategy two steps. We train network semi-synthetic...

This paper proposes a general framework to solve Non-Rigid Shape-from-Motion (NRSfM) with the perspective camera under isometric deformations. Contrary usual low-rank linear shape basis, isometry allows us recover complex deformations from sparse set of images. Existing methods suffer ambiguities and may be very expensive solve. We bring four main contributions. First, we formulate NRSfM as system first-order Partial Differential Equations (PDE) involving shape’s depth normal field an...

Shape-from-Template (SfT) reconstructs the shape of a deforming surface from single image, 3D template and deformation prior. For isometric deformations, this is well-posed problem. However, previous methods which require no initialization break down when perspective effects are small, happens object small or viewed larger distances. That is, they do not handle all projection geometries. We propose stable SfT that accurately reconstruct for follow existing approach using first-order...

An active research objective in Computer Assisted Intervention (CAI) is to develop guidance systems aid surgical teams laparoscopic Minimal Invasive Surgery (MIS) using Augmented Reality (AR). This involves registering and fusing additional data from other modalities overlaying it onto the video realtime. We present first AR-based image system for assisted myoma localisation uterine laparosurgery. a framework semi-automatically pre-operative Magnetic Resonance Image (MRI) with deformable...

This paper presents a sensor system for robot localization based on the information obtained from single camera attached in fixed place external to robot. Our approach firstly obtains 3D geometrical model of projection its natural appearance while performs an initialization trajectory. proposes structure-from-motion solution that uses odometry sensors inside as metric reference. Secondly, online method sequential Bayesian inference is proposed, which link between image measurements and pose...

It has been recently shown that reconstructing an isometric surface from a single 2D input image matched to 3D template was well-posed problem. This however does not tell us how reconstruction algorithms will behave in practical conditions, where the amount of perspective is generally small and projection thus behaves like weak-perspective or orthography. We here bring answers what theoretically recoverable such imaging explain why existing convex numerical solutions analytical may be...

This paper presents an algorithm that automatically corrects the distortion caused by multipath interference (MpI) in depth measurements obtained with time of flight cameras (ToF cameras). A radiometric model explains, under some mild simplifications, working principle a ToF camera including for MpI is proposed. Using this we demonstrate all information needed compensating influence on scene captured self-contained (depth and amplitude infrared signal). We propose iterative optimization...

We present a global and convex formulation for the template-less 3D reconstruction of deforming object with perspective camera. show first time how to construct Second-Order Cone Programming (SOCP) problem Non-Rigid Structure-from-Motion (NRSfM) using Maximum-Depth Heuristic (MDH). In this regard, we deviate strongly from general trend affine cameras factorization-based methods solve NRSfM, which do not perform well complex nonlinear deformations. MDH, points' depths are maximized so that...

We study Isometric Non-Rigid Shape-from-Motion (Iso-NRSfM): given multiple intrinsically calibrated monocular images, we want to reconstruct the time-varying 3D shape of an object undergoing isometric deformations. show that Iso-NRSfM is solvable from warps (the inter-image geometric transformations). propose a new theoretical framework based on Riemmanian manifolds represent unknown surfaces, as embeddings camera's retinal planes. This allows us use manifolds' metric tensor and Christoffel...