Abstract: In this paper we present a method for document categorization which processes incoming document images such as invoices or receipts. The categorization of these document images is done in terms of the presence of a certain graphical logo detected without segmentation. The graphical logos are described by a set of local features and the categorization of the documents is performed by the use of a bag-of-words model. Spatial coherence rules are added to reinforce the correct category hypothesis, aiming also to spot the logo inside the document image. Experiments which demonstrate the effectiveness of this system on a large set of real data are presented.

Abstract: Handwritten marriage licenses books are characterized by a simple structure of the text in the records with an evolutionary vocabulary, mainly composed of proper names that change along the time. This distinct vocabulary makes automatic transcription and semantic information extraction difficult tasks. Previous works have shown that the use of category-based language models and a Grammatical Inference technique known as MGGI can improve the accuracy of these
tasks. However, the application of the MGGI algorithm requires an a priori knowledge to label the words of the training strings, that is not always easy to obtain. In this paper we study how to automatically obtain the information required by the MGGI algorithm using a technique based on Confusion Networks. Using the resulting language model, full handwritten text recognition and information extraction experiments have been carried out with results supporting the proposed approach.

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: Activity recognition from long unstructured egocentric photo-streams has several applications in assistive technology such as health monitoring and frailty detection, just to name a few. However, one of its main technical challenges is to deal with the low frame rate of wearable photo-cameras, which causes abrupt appearance changes between consecutive frames. In consequence, important discriminatory low-level features from motion such as optical flow cannot be estimated. In this paper, we present a batch-driven approach for training a deep learning architecture that strongly rely on Long short-term units to tackle this problem. We propose two different implementations of the same approach that process a photo-stream sequence using batches of fixed size with the goal of capturing the temporal evolution of high-level features. The main difference between these implementations is that one explicitly models consecutive batches by overlapping them. Experimental results over a public dataset acquired by three users demonstrate the validity of the proposed architectures to exploit the temporal evolution of convolutional features over time without relying on event boundaries.

Abstract: This work aims at detecting in which regions the objects in the image are by using information about the intensity of valleys, which appear to surround ob- jects in images where the source of light is in the line of direction than the camera. We present our depth of valleys accumulation method, which consists of two stages: first, the definition of the depth of valleys image which combines the output of a ridges and valleys detector with the morphological gradient to measure how deep is a point inside a valley and second, an algorithm that denotes points of the image as interior to objects those which are inside complete or incomplete boundaries in the depth of valleys image. To evaluate the performance of our method we have tested it on several application domains. Our results on object region identification are promising, specially in the field of polyp detection in colonoscopy videos, and we also show its applicability in different areas.

Abstract: XBadges is a research project based on the hypothesis that commercial video games (nonserious games) can train soft skills. We measure persistence, patial reasoning and risk taking before and after subjects paticipate in controlled game playing sessions.
In addition, we have developed an automatic facial expression recognition system capable of inferring their emotions while playing, allowing us to study the role of emotions in soft skills acquisition. We have used Flappy Bird, Pacman and Tetris for assessing changes in persistence, risk taking and spatial reasoning respectively.
Results show how playing Tetris significantly improves spatial reasoning and how playing Pacman significantly improves prudence in certain areas of behavior. As for emotions, they reveal that being concentrated helps to improve performance and skills acquisition. Frustration is also shown as a key element. With the results obtained we are able to glimpse multiple applications in areas which need soft skills development.

Abstract: Intravascular ultrasound images represent a unique tool to analyze the morphology of arteries and vessels (plaques, restenosis, etc). The poor quality of these images makes unsupervised segmentation based on traditional segmentation algorithms (such as edge or ridge/valley detection) fail to achieve the expected results. In this paper we present a probabilistic flexible template to separate different regions in the image. In particular, we use elliptic templates to model and detect the shape of the vessel inner wall in IVUS images. We present the results of successful segmentation obtained from patients undergoing stent treatment. A physician team has validated these results.