Abstract: Deep learning has recently been enjoying an increasing popularity due to its success in solving challenging tasks. In particular, deep learning has proven to be effective in a large variety of computer vision tasks, such as image classification, object recognition and image parsing. Contrary to previous research, which required engineered feature representations, designed by experts, in order to succeed, deep learning attempts to learn representation hierarchies automatically from data. More recently, the trend has been to go deeper with representation hierarchies.
Learning (very) deep representation hierarchies is a challenging task, which
involves the optimization of highly non-convex functions. Therefore, the search
for algorithms to ease the learning of (very) deep representation hierarchies from data is extensive and ongoing.
In this thesis, we tackle the challenging problem of easing the learning of (very) deep representation hierarchies. We present a hyper-parameter free, off-the-shelf, simple and fast unsupervised algorithm to discover hidden structure from the input data by enforcing a very strong form of sparsity. We study the applicability and potential of the algorithm to learn representations of varying depth in a handful of applications and domains, highlighting the ability of the algorithm to provide discriminative feature representations that are able to achieve top performance.
Yet, while emphasizing the great value of unsupervised learning methods when
labeled data is scarce, the recent industrial success of deep learning has revolved around supervised learning. Supervised learning is currently the focus of many recent research advances, which have shown to excel at many computer vision tasks. Top performing systems often involve very large and deep models, which are not well suited for applications with time or memory limitations. More in line with the current trends, we engage in making top performing models more efficient, by designing very deep and thin models. Since training such very deep models still appears to be a challenging task, we introduce a novel algorithm that guides the training of very thin and deep models by hinting their intermediate representations.
Very deep and thin models trained by the proposed algorithm end up extracting feature representations that are comparable or even better performing
than the ones extracted by large state-of-the-art models, while compellingly
reducing the time and memory consumption of the model.

Abstract: Handwritten word-spotting is traditionally viewed as an image matching task between one or multiple query word-images and a set of candidate word-images in a database. This is a typical instance of the query-by-example paradigm. In this article, we introduce a statistical framework for the word-spotting problem which employs hidden Markov models (HMMs) to model keywords and a Gaussian mixture model (GMM) for score normalization. We explore the use of two types of HMMs for the word modeling part: continuous HMMs (C-HMMs) and semi-continuous HMMs (SC-HMMs), i.e. HMMs with a shared set of Gaussians. We show on a challenging multi-writer corpus that the proposed statistical framework is always superior to a traditional matching system which uses dynamic time warping (DTW) for word-image distance computation. A very important finding is that the SC-HMM is superior when labeled training data is scarce—as low as one sample per keyword—thanks to the prior information which can be incorporated in the shared set of Gaussians.

Abstract: This paper presents a novel formulation, which derives in a smooth minimization problem, to tackle the rigid registration between a given point set and a model set. Unlike most of the existing works, which are based on minimizing a point-wise correspondence term, we propose to describe the model set by means of an implicit representation. It allows a new definition of the registration error, which works beyond the point level representation. Moreover, it could be used in a gradient-based optimization framework. The proposed approach consists of two stages. Firstly, a novel formulation is proposed that relates the registration parameters with the distance between the model and data set. Secondly, the registration parameters are obtained by means of the Levengberg-Marquardt algorithm. Experimental results and comparisons with state of the art show the validity of the proposed framework.

Abstract: Impact Factor
This paper presents a simple distance estimation for implicit polynomial fitting. It is computed as the height of a simplex built between the point and the surface (i.e., a triangle in 2-D or a tetrahedron in 3-D), which is used as a coarse but reliable estimation of the orthogonal distance. The proposed distance can be described as a function of the coefficients of the implicit polynomial. Moreover, it is differentiable and has a smooth behavior . Hence, it can be used in any gradient-based optimization. In this paper, its use in a Levenberg-Marquardt framework is shown, which is particularly devoted for nonlinear least squares problems. The proposed estimation is a generalization of the gradient-based distance estimation, which is widely used in the literature. Experimental results, both in 2-D and 3-D data sets, are provided. Comparisons with state-of-the-art techniques are presented, showing the advantages of the proposed approach.

Abstract: In this paper, the registration problem is formulated as a point to model distance minimization. Unlike most of the existing works, which are based on minimizing a point-wise correspondence term, this formulation avoids the correspondence search that is time-consuming. In the first stage, the target set is described through an implicit function by employing a linear least squares fitting. This function can be either an implicit polynomial or an implicit B-spline from a coarse to fine representation. In the second stage, we show how the obtained implicit representation is used as an interface to convert point-to-point registration into point-to-implicit problem. Furthermore, we show that this registration distance is smooth and can be minimized through the Levengberg-Marquardt algorithm. All the formulations presented for both stages are compact and easy to implement. In addition, we show that our registration method can be handled using any implicit representation though some are coarse and others provide finer representations; hence, a tradeoff between speed and accuracy can be set by employing the right implicit function. Experimental results and comparisons in 2D and 3D show the robustness and the speed of convergence of the proposed approach.