This thesis presents new methods in two closely related areas of computer vision: human pose estimation, and gesture recognition in videos. In human pose estimation, we show that random forests can be used to estimate human pose in monocular videos. To this end, we propose a co-segmentation algorithm for segmenting humans out of videos, and an evaluator that predicts whether the estimated poses...

We are going to introduce the concept of stochastic 3D modeling of geometrically complex (disordered) microstructures as a tool for virtual materials testing, including applications to battery electrodes as well as to electrodes of fuel cells, and solar cells. Using stochastic 3D models, one can generate a large variety of stochastically simulated micrsotructures with little computational effor...

High spectral dimensionality and the shortage of annotations make hyperspectral image (HSI) classification a challenging problem. Recent studies suggest that convolutional neural networks can learn discriminative spatial features, which play a paramount role in HSI interpretation. However, most of these methods ignore the distinctive spectral-spatial characteristic of hyperspectral data. In add...

We propose a dual pathway, 11-layers deep, three-dimensional Convolutional Neural Network for the challenging task of brain lesion segmentation. The devised architecture is the result of an in-depth analysis of the limitations of current networks proposed for similar applications. To overcome the computational burden of processing 3D medical scans, we have devised an efficient and effective den...

Inspired by the recent success of methods that employ shape priors to achieve robust 3D reconstructions, we propose a novel recurrent neural network architecture that we call the 3D Recurrent Reconstruction Neural Network (3D-R2N2). The network learns a mapping from images of objects to their underlying 3D shapes from a large collection of synthetic data [13]. Our network takes in one or more i...

We present a method to reconstruct a disordered network of thin biopolymers, such as collagen gels, from three-dimensional (3D) image stacks recorded with a confocal microscope. The method is based on a template matching algorithm that simultaneously performs a binarization and skeletonization of the network. The size and intensity pattern of the template is automatically adapted to the input d...

We present a numerically efficient method to reconstruct a disordered network of thin biopolymers, such as collagen gels, from three-dimensional (3D) image stacks recorded with a confocal microscope. Our method is based on a template matching algorithm that simultaneously performs a binarization and skeletonization of the network. The size and intensity pattern of the template is automatically ...

Human can obtain 3D percept from 2D images on the retinas. However, the neural mechanism of biological stereo vision is largely unknown. On the other hand, Markov random field and belief propagation algorithm have given birth to stereo algorithms with top performance. I propose that brain may employ similar neural network to perceive depth. I design a model combining Markov network with known p...

3D reconstruction applications can benefit greatly from knowledge about coplanar feature points. Extracting this knowledge from images alone is a difficult task, however. The typical approach to this problem is to search for homographies in a set of point correspondences using the RANSAC algorithm. In this work we focus on two open issues with a blind random search. First, we enforce the detect...