Abstract: The representation (or feature) learning has been an emerging concept in the last years, since it collects a set of techniques that are present in any theoretical or practical methodology referring to artificial intelligence. In computer vision, a very common representation has adopted the form of the well-known Bag of Visual Words. This representation appears implicitly in most approaches where images are described, and is also present in a huge number of areas and domains: image content retrieval, pedestrian detection, human-computer interaction, surveillance, e-health, and social computing, amongst others. The early stages of this dissertation provide an approach for learning visual representations inside evolutionary algorithms, which consists of evolving weighting schemes to improve the BoVW representations for the task of recognizing categories of videos and images. Thus, we demonstrate the applicability of the most common weighting schemes, which are often used in text mining but are less frequently found in computer vision tasks. Beyond learning these visual representations, we provide an approach based on fusion strategies for learning spatiotemporal representations, from multimodal data obtained by depth sensors. Besides, we specially aim at the evolutionary and dynamic modelling, where the temporal factor is present in the nature of the data, such as video sequences of gestures and actions. Indeed, we explore the effects of probabilistic modelling for those approaches based on dynamic programming, so as to handle the temporal deformation and variance amongst video sequences of different categories. Finally, we integrate dynamic programming and generative models into an evolutionary computation framework, with the aim of learning Bags of SubGestures (BoSG) representations and hence to improve the generalization capability of standard gesture recognition approaches. The results obtained in the experimentation demonstrate, first, that evolutionary algorithms are useful for improving the representation of BoVW approaches in several datasets for recognizing categories in still images and video sequences. On the other hand, our experimentation reveals that both, the use of dynamic programming and generative models to align video sequences, and the representations obtained from applying fusion strategies in multimodal data, entail an enhancement on the performance when recognizing some gesture categories. Furthermore, the combination of evolutionary algorithms with models based on dynamic programming and generative approaches results, when aiming at the classification of video categories on large video datasets, in a considerable improvement over standard gesture and action recognition approaches. Finally, we demonstrate the applications of these representations in several domains for human analysis: classification of images where humans may be present, action and gesture recognition for general applications, and in particular for conversational settings within the field of restorative justice

Abstract: In this work we present a novel compact scene representation based on Stixels that infers geometric and semantic information. Our approach overcomes the previous rather restrictive geometric assumptions for Stixels by introducing a novel depth model to account for non-flat roads and slanted objects. Both semantic and depth cues are used jointly to infer the scene representation in a sound global energy minimization formulation. Furthermore, a novel approximation scheme is introduced that uses an extremely efficient over-segmentation. In doing so, the computational complexity of the Stixel inference algorithm is reduced significantly, achieving real-time computation capabilities with only a slight drop in accuracy. We evaluate the proposed approach in terms of semantic and geometric accuracy as well as run-time on four publicly available benchmark datasets. Our approach maintains accuracy on flat road scene datasets while improving substantially on a novel non-flat road dataset.

Abstract: Natural image modeling is a landmark challenge of unsupervised learning. Variational Autoencoders (VAEs) learn a useful latent representation and generate samples that preserve global structure but tend to suffer from image blurriness. PixelCNNs model sharp contours and details very well, but lack an explicit latent representation and have difficulty modeling large-scale structure in a computationally efficient way. In this paper, we present PixelVAE, a VAE model with an autoregressive decoder based on PixelCNN. The resulting architecture achieves state-of-the-art log-likelihood on binarized MNIST. We extend PixelVAE to a hierarchy of multiple latent variables at different scales; this hierarchical model achieves competitive likelihood on 64x64 ImageNet and generates high-quality samples on LSUN bedrooms.

Abstract: Automated video and image analysis can be a very efficient tool to analyze
animal behavior based on sociality, especially in hard access environments
for researchers. The understanding of this social behavior can play a key role in the sustainable design of capture policies of many species. This paper proposes the use of computer vision algorithms to identify and track a specific specie, the Norway lobster, Nephrops norvegicus, a burrowing decapod with relevant commercial value which is captured by trawling. These animals can only be captured when are engaged in seabed excursions, which are strongly related with their social behavior.
This emergent behavior is modulated by the day-night cycle, but their social
interactions remain unknown to the scientific community. The paper introduces an identification scheme made of four distinguishable black and white tags (geometric shapes). The project has recorded 15-day experiments in laboratory pools, under monochromatic blue light (472 nm.) and darkness conditions (recorded using Infra Red light). Using this massive image set, we propose a comparative of state-ofthe-art computer vision algorithms to distinguish and track the different animals’ movements. We evaluate the robustness to the high noise presence in the infrared video signals and free out-of-plane rotations due to animal movement. The experiments show promising accuracies under a cross-validation protocol, being adaptable to the automation and analysis of large scale data. In a second contribution, we created an extensive dataset of shapes (46027 different shapes) from four daily experimental video recordings, which will be available to the community.

Abstract: It has recently been demonstrated that local feature descriptors based on convolutional neural networks (CNN) can significantly improve the matching performance. Previous work on learning such descriptors has focused on exploiting pairs of positive and negative patches to learn discriminative CNN representations. In this work, we propose to utilize triplets of training samples, together with in-triplet mining of hard negatives.
We show that our method achieves state of the art results, without the computational overhead typically associated with mining of negatives and with lower complexity of the network architecture. We compare our approach to recently introduced convolutional local feature descriptors, and demonstrate the advantages of the proposed methods in terms of performance and speed. We also examine different loss functions associated with triplets.

Abstract: Text Proposals have emerged as a class-dependent version of object proposals – efficient approaches to reduce the search space of possible text object locations in an image. Combined with strong word classifiers, text proposals currently yield top state of the art results in end-to-end scene text
recognition. In this paper we propose an improvement over the original Text Proposals algorithm of [1], combining it with Fully Convolutional Networks to improve the ranking of proposals. Results on the ICDAR RRC and the COCO-text datasets show superior performance over current state-of-the-art.