Manuel Marques

Optimal shape from motion estimation with missing and degenerate data

Reconstructing a 3D scene from a moving camera is one of the most important issues in the field of computer vision. In this scenario, not all points are known in all images (e.g. due to occlusion), thus generating missing data. The state of the art handles the missing points in this context by enforcing rank constraints on the point track matrix. However, quite frequently, close up views tend to capture planar surfaces producing degenerate data. If one single frame is degenerate, the whole sequence will produce high errors on the shape reconstruction, even though the observation matrix verifies the rank 4 constraint. In this paper, we propose to solve the structure from motion problem with degenerate data, introducing a new factorization algorithm that imposes the full scaled orthographic model in one single optimization procedure. By imposing all model constraints, a unique (correct) 3D shape is estimated regardless of the data degeneracies. Experiments show that remarkably good reconstructions are obtained with an approximate models such as orthography.

The use of dual-stream functions in the analysis of three-dimensional metal forming processes

Abstract Modelling metal forming processes with the dual stream function technique is considered. The dual stream function concept is reviewed. Its numerical implementation in a computer program is presented. The method enables the numerical evaluation of the process power and gives detailed information about other process parameters such as friction between the workpiece and the tool, workpiece geometry and material characteristics. Two selected examples concerning three-dimensional flat rolling and polygonal forging are analysed by this technique. The theoretical predictions are compared with results obtained by other authors using alternative methods. Computational advantages over other methods, namely the finite element method are the main feature of the present analysis.

Subspace matching: Unique solution to point matching with geometric constraints

Finding correspondences between feature points is one of the most relevant problems in the whole set of visual tasks. In this paper we address the problem of matching a feature vector (or a matrix) to a given subspace. Given any vector base of such a subspace, we observe a linear combination of its elements with all entries swapped by an unknown permutation. We prove that such a computationally hard integer problem is uniquely solved in a convex set resulting from relaxing the original problem. Also, if noise is present, based on this result, we provide a robust estimate recurring to a linear programming-based algorithm. We use structure-from-motion and object recognition as motivating examples.

Monitoring the wear of sintered high-speed-steel tools

Abstract This paper presents the results of a detailed investigation into the relationship between the wear of sintered T15 high-speed-steel cutting tools and the associated cutting forces. The equations for the cutting forces are established taking into account the influence of both crater and flank wear. The experimental results show that during the initial stage of wear, crater wear promotes a decrease of the cutting forces. This behaviour of the cutting forces is related to the change of the actual side-rake angle, which depends on the crater depth and width. When the true rake-angle reaches a maximum, the cutting forces tend to stabilize, unless flank wear progresses at a high wear-rate, in which case the cutting forces increase until the failure of the cutting tool. According to the proposed model, the measurement of cutting forces could provide an adequate control variable for monitoring the wear of high-speed-steel tools.

A solution to plane strain extrusion by the upper bound approach and the weighted residuals method

Abstract The analysis of plane strain extrusion through wedge shaped dies is based in a general kinematically admissible velocity field coupled with the method of weighted residuals. The expressions of velocity and strain rate are derived analytically and the resulting fields for several extrusion ratios are obtained by minimizing the upper bound power with respect to free parameters introduced in the inlet and outlet boundaries. The stresses are calculated from a stress function, using the strain rate field and the boundary conditions, by the least squares method. The results are compared with the complete solution obtained by the slip line field. The technique is shown to give satisfactory results and can be applied to more complex metalforming processes, as it requires only modest computing time and resources.

3D face recognition from multiple images: A shape-from-motion approach

In this article we explore the use of methodologies for 3D reconstruction from multiple images to recognize faces. We try to devise a strategy to tackle the problem of recognizing faces from images exhibiting strong pose (rotation and occlusion) and without prior knowledge (uncalibrated cameras, images from different sources). We do so by framing the recognition in the context of 3D structure from motion with missing data problems. In fact, recently, there has been a strong trend towards using 3D information to verify and recognize faces. However most of the state of the art works are developed over 3D sensors (3D range finders, stereo).Here we propose to do recognition by measuring the likelihood that one or more images were generated by a given 3D shape. In other words, given the subjects 3D shape, we compute the required transformation to match the probe images and measure its deviation from rigidity. This process is not straightforward due to the strong pose: face points are not visible in all images so there is a ldquomissing datardquo problem intrinsic to the formulation. One key assumption is that all images (and shape) correspond to neutral expression. Though limited in scale due to lack of large databases, a set of tests demonstrate the adequacy and good performance of the approach.

Lamp: Linear approach for matching points

The 3D object recognition from a single or multiple 2D images is a very important problem in the computer vision field with a wide range of real applications. Considering the affine camera model, the main issue in solving this problem is the matching process between the objects 3D points and their 2D projections. In this work, we tackle the 3D-2D matching problem. It is formulated as a finite set of independent linear programs, solved efficiently. The 2D-2D and 3D-3D are also discussed. To show the validity of the proposed method, synthetic and real experiments are performed.

Analysis of three-dimensional hexagonal closed-die heading by the finite-element flow formulation

Abstract Modelling three-dimensional metal-forming processes with the finite-element flow formulation based on the Doraivelu et al. constitutive equations is considered. The basic theory is reviewed. Special emphasis is placed on the numerical implementation of such formulation. The authors discuss the substitution of the Gauss by the Irons numerical integration rule for the evaluation of the deviatoric component of the generalized stiffness matrix, and present two criteria for analysing mesh degeneracy and mesh-die interference. Three-dimensional contact problems are discussed briefly. Three-dimensional closed-die hexagonal heading, starting from a cylindrical aluminium-alloy billet, is reported. Theoretical and experimental results for the punch load-punch displacement evolution are analysed, good agreement between them being found.

SMARTcycling: Assessing cyclists' driving experience

Due to economic and environmental issues, bicycles have been regaining their significance as a transportation vehicle in urban scenarios. To further drive this desirable trend, policy makers must have the tools to access current bicycle infrastructures and road safety concerns. Fundamental for this assessment is a deeper understanding of how cyclists use current infrastructures, if the cycling experience results in stressful events, and the conditions of the current infrastructure. We here introduce a new platform, SMARTcycling, that, by taking advantage of the mobile power available to a smartphone, captures and stores data from several sensors, namely an action camera, a cardio signal acquisition belt, and smartphones Global Positioning System (GPS) coordinates. The data is further processed and, through visual cues, we access the cyclist driving events and road condition cues. SMARTcycling also detects the cyclist stress using the electrocardiograms (ECG) from the belt. We further contribute by making available a dataset containing the sensors data from 10 paths over two cities in Portugal. On this dataset, we show our initial promising results on event detection, road condition identification and stress assessment.

Hot forging of shells: An upper-bound approach

Abstract An analytical method is proposed for estimating the load in the axisymmetric backward trusion/piercing of tubular components. The method is based on a general kinematically admissible velocity field. For this purpose the geometry is previously split in several sub-domains, the configuration of the velocity discontinuity surfaces between them being determined by minimizing the upper-bound power with respect to free parameters. The theoretical predictions of forming force for a 120 mm artillery shell are compared with the results obtained under production conditions, such comparison indicating that this method can predict the forming load quite satisfactorily, with the further advantage that it can be implemented easily within an industrial environment since the method requires only the use of a personal computer.