General trajectory prior for non-rigid reconstruction

Trajectory basis Non-Rigid Structure From Motion (NRSFM) currently faces two problems: the limit of reconstructability and the need to tune the basis size for different sequences. This paper provides a novel theoretical bound on 3D reconstruction error, arguing that the existing definition of reconstructability is fundamentally flawed in that it fails to consider system condition. This insight motivates a novel strategy whereby the trajectory's response to a set of high-pass filters is minimised. The new approach eliminates the need to tune the basis size and is more efficient for long sequences. Additionally, the truncated DCT basis is shown to have a dual interpretation as a high-pass filter. The success of trajectory filter reconstruction is demonstrated quantitatively on synthetic projections of real motion capture sequences and qualitatively on real image sequences.

Multi-modal deformable medical image registration

Non-rigid image registration is an essential tool required for overcoming the inherent local anatomical variations that exist between images acquired from different individuals or atlases. Furthermore, certain applications require this type of registration to operate across images acquired from different imaging modalities. One popular local approach for estimating this registration is a block matching procedure utilising the mutual information criterion. However, previous block matching procedures generate a sparse deformation field containing displacement estimates at uniformly spaced locations. This neglects to make use of the evidence that block matching results are dependent on the amount of local information content. This paper presents a solution to this drawback by proposing the use of a Reversible Jump Markov Chain Monte Carlo statistical procedure to optimally select grid points of interest. Three different methods are then compared to propagate the estimated sparse deformation field to the entire image including a thin-plate spline warp, Gaussian convolution, and a hybrid fluid technique. Results show that non-rigid registration can be improved by using the proposed algorithm to optimally select grid points of interest.

Locally weighted learning methods for non-rigid robot control

This thesis develops a novel approach to robot control that learns to account for a robot's dynamic complexities while executing various control tasks using inspiration from biological sensorimotor control and machine learning. A robot that can learn its own control system can account for complex situations and adapt to changes in control conditions to maximise its performance and reliability in the real world. This research has developed two novel learning methods, with the aim of solving issues with learning control of non-rigid robots that incorporate additional dynamic complexities. The new learning control system was evaluated on a real three degree-of-freedom elastic joint robot arm with a number of experiments: initially validating the learning method and testing its ability to generalise to new tasks, then evaluating the system during a learning control task requiring continuous online model adaptation.

Improving fluid registration through white matter segmentation in a twin study design

Robust and automatic non-rigid registration depends on many parameters that have not yet been systematically explored. Here we determined how tissue classification influences non-linear fluid registration of brain MRI. Twin data is ideal for studying this question, as volumetric correlations between corresponding brain regions that are under genetic control should be higher in monozygotic twins (MZ) who share 100% of their genes when compared to dizygotic twins (DZ) who share half their genes on average. When these substructure volumes are quantified using tensor-based morphometry, improved registration can be defined based on which method gives higher MZ twin correlations when compared to DZs, as registration errors tend to deplete these correlations. In a study of 92 subjects, higher effect sizes were found in cumulative distribution functions derived from statistical maps when performing tissue classification before fluid registration, versus fluidly registering the raw images. This gives empirical evidence in favor of pre-segmenting images for tensor-based morphometry.

Closed-form solutions for low-rank non-rigid reconstruction

Recovering the motion of a non-rigid body from a set of monocular images permits the analysis of dynamic scenes in uncontrolled environments. However, the extension of factorisation algorithms for rigid structure from motion to the low-rank non-rigid case has proved challenging. This stems from the comparatively hard problem of finding a linear “corrective transform” which recovers the projection and structure matrices from an ambiguous factorisation. We elucidate that this greater difficulty is due to the need to find multiple solutions to a non-trivial problem, casting a number of previous approaches as alleviating this issue by either a) introducing constraints on the basis, making the problems nonidentical, or b) incorporating heuristics to encourage a diverse set of solutions, making the problems inter-dependent. While it has previously been recognised that finding a single solution to this problem is sufficient to estimate cameras, we show that it is possible to bootstrap this partial solution to find the complete transform in closed-form. However, we acknowledge that our method minimises an algebraic error and is thus inherently sensitive to deviation from the low-rank model. We compare our closed-form solution for non-rigid structure with known cameras to the closed-form solution of Dai et al. [1], which we find to produce only coplanar reconstructions. We therefore make the recommendation that 3D reconstruction error always be measured relative to a trivial reconstruction such as a planar one.

Stationary processes for object detection and non-rigid structure-from-motion

Stationary processes are random variables whose value is a signal and whose distribution is invariant to translation in the domain of the signal. They are intimately connected to convolution, and therefore to the Fourier transform, since the covariance matrix of a stationary process is a Toeplitz matrix, and Toeplitz matrices are the expression of convolution as a linear operator. This thesis utilises this connection in the study of i) efficient training algorithms for object detection and ii) trajectory-based non-rigid structure-from-motion.

A subject-specific model of human buttocks and thighs in a seated posture

Finite element analyses of the human body in seated postures requires digital models capable of providing accurate and precise prediction of the tissue-level response of the body in the seated posture. To achieve such models, the human anatomy must be represented with high fidelity. This information can readily be defined using medical imaging techniques such as Magnetic Resonance Imaging (MRI) or Computed Tomography (CT). Current practices for constructing digital human models, based on the magnetic resonance (MR) images, in a lying down (supine) posture have reduced the error in the geometric representation of human anatomy relative to reconstructions based on data from cadaveric studies. Nonetheless, the significant differences between seated and supine postures in segment orientation, soft-tissue deformation and soft tissue strain create a need for data obtained in postures more similar to the application posture. In this study, we present a novel method for creating digital human models based on seated MR data. An adult-male volunteer was scanned in a simulated driving posture using a FONAR 0.6T upright MRI scanner with a T1 scanning protocol. To compensate for unavoidable image distortion near the edges of the study, images of the same anatomical structures were obtained in transverse and sagittal planes. Combinations of transverse and sagittal images were used to reconstruct the major anatomical features from the buttocks through the knees, including bone, muscle and fat tissue perimeters, using Solidworks® software. For each MR image, B-splines were created as contours for the anatomical structures of interest, and LOFT commands were used to interpolate between the generated Bsplines. The reconstruction of the pelvis, from MR data, was enhanced by the use of a template model generated in previous work CT images. A non-rigid registration algorithm was used to fit the pelvis template into the MR data. Additionally, MR image processing was conducted to both the left and the right sides of the model due to the intended asymmetric posture of the volunteer during the MR measurements. The presented subject-specific, three-dimensional model of the buttocks and thighs will add value to optimisation cycles in automotive seat development when used in simulating human interaction with automotive seats.

Rank minimization across appearance and shape for AAM ensemble fitting

Active Appearance Models (AAMs) employ a paradigm of inverting a synthesis model of how an object can vary in terms of shape and appearance. As a result, the ability of AAMs to register an unseen object image is intrinsically linked to two factors. First, how well the synthesis model can reconstruct the object image. Second, the degrees of freedom in the model. Fewer degrees of freedom yield a higher likelihood of good fitting performance. In this paper we look at how these seemingly contrasting factors can complement one another for the problem of AAM fitting of an ensemble of images stemming from a constrained set (e.g. an ensemble of face images of the same person).

Nonrigid face alignment for unknown subject in video

Non-rigid face alignment is a very important task in a large range of applications but the existing tracking based non-rigid face alignment methods are either inaccurate or requiring person-specific model. This dissertation has developed simultaneous alignment algorithms that overcome these constraints and provide alignment with high accuracy, efficiency, robustness to varying image condition, and requirement of only generic model.

A new registration method based on Log-Euclidean tensor metrics and its application to genetic studies

In structural brain MRI, group differences or changes in brain structures can be detected using Tensor-Based Morphometry (TBM). This method consists of two steps: (1) a non-linear registration step, that aligns all of the images to a common template, and (2) a subsequent statistical analysis. The numerous registration methods that have recently been developed differ in their detection sensitivity when used for TBM, and detection power is paramount in epidemological studies or drug trials. We therefore developed a new fluid registration method that computes the mappings and performs statistics on them in a consistent way, providing a bridge between TBM registration and statistics. We used the Log-Euclidean framework to define a new regularizer that is a fluid extension of the Riemannian elasticity, which assures diffeomorphic transformations. This regularizer constrains the symmetrized Jacobian matrix, also called the deformation tensor. We applied our method to an MRI dataset from 40 fraternal and identical twins, to revealed voxelwise measures of average volumetric differences in brain structure for subjects with different degrees of genetic resemblance.

Geometry vs appearance for discriminating between posed and spontaneous emotions

Spontaneous facial expressions differ from posed ones in appearance, timing and accompanying head movements. Still images cannot provide timing or head movement information directly. However, indirectly the distances between key points on a face extracted from a still image using active shape models can capture some movement and pose changes. This information is superposed on information about non-rigid facial movement that is also part of the expression. Does geometric information improve the discrimination between spontaneous and posed facial expressions arising from discrete emotions? We investigate the performance of a machine vision system for discrimination between posed and spontaneous versions of six basic emotions that uses SIFT appearance based features and FAP geometric features. Experimental results on the NVIE database demonstrate that fusion of geometric information leads only to marginal improvement over appearance features. Using fusion features, surprise is the easiest emotion (83.4% accuracy) to be distinguished, while disgust is the most difficult (76.1%). Our results find different important facial regions between discriminating posed versus spontaneous version of one emotion and classifying the same emotion versus other emotions. The distribution of the selected SIFT features shows that mouth is more important for sadness, while nose is more important for surprise, however, both the nose and mouth are important for disgust, fear, and happiness. Eyebrows, eyes, nose and mouth are important for anger.

Graph rigidity for near-coplanar structure from motion

Recent algorithms for monocular motion capture (MoCap) estimate weak-perspective camera matrices between images using a small subset of approximately-rigid points on the human body (i.e. the torso and hip). A problem with this approach, however, is that these points are often close to coplanar, causing canonical linear factorisation algorithms for rigid structure from motion (SFM) to become extremely sensitive to noise. In this paper, we propose an alternative solution to weak-perspective SFM based on a convex relaxation of graph rigidity. We demonstrate the success of our algorithm on both synthetic and real world data, allowing for much improved solutions to marker less MoCap problems on human bodies. Finally, we propose an approach to solve the two-fold ambiguity over bone direction using a k-nearest neighbour kernel density estimator.

Fourier Lucas-Kanade Algorithm

In this paper we propose a framework for both gradient descent image and object alignment in the Fourier domain. Our method centers upon the classical Lucas & Kanade (LK) algorithm where we represent the source and template/model in the complex 2D Fourier domain rather than in the spatial 2D domain. We refer to our approach as the Fourier LK (FLK) algorithm. The FLK formulation is advantageous when one pre-processes the source image and template/model with a bank of filters (e.g. oriented edges, Gabor, etc.) as: (i) it can handle substantial illumination variations, (ii) the inefficient pre-processing filter bank step can be subsumed within the FLK algorithm as a sparse diagonal weighting matrix, (iii) unlike traditional LK the computational cost is invariant to the number of filters and as a result far more efficient, and (iv) this approach can be extended to the inverse compositional form of the LK algorithm where nearly all steps (including Fourier transform and filter bank pre-processing) can be pre-computed leading to an extremely efficient and robust approach to gradient descent image matching. Further, these computational savings translate to non-rigid object alignment tasks that are considered extensions of the LK algorithm such as those found in Active Appearance Models (AAMs).