Abstract: This paper investigates the role of saliency to improve the classification accuracy of a Convolutional Neural Network (CNN) for the case when scarce training data is available. Our approach consists in adding a saliency branch to an existing CNN architecture which is used to modulate the standard bottom-up visual features from the original image input, acting as an attentional mechanism that guides the feature extraction process. The main aim of the proposed approach is to enable the effective training of a fine-grained recognition model with limited training samples and to improve the performance on the task, thereby alleviating the need to annotate a large dataset. The vast majority of saliency methods are evaluated on their ability to generate saliency maps, and not on their functionality in a complete vision pipeline. Our proposed pipeline allows to evaluate saliency methods for the high-level task of object recognition. We perform extensive experiments on various fine-grained datasets (Flowers, Birds, Cars, and Dogs) under different conditions and show that saliency can considerably improve the network’s performance, especially for the case of scarce training data. Furthermore, our experiments show that saliency methods that obtain improved saliency maps (as measured by traditional saliency benchmarks) also translate to saliency methods that yield improved performance gains when applied in an object recognition pipeline.

Abstract: For many applications the collection of labeled data is expensive laborious. Exploitation of unlabeled data during training is thus a long pursued objective of machine learning. Self-supervised learning addresses this by positing an auxiliary task (different, but related to the supervised task) for which data is abundantly available. In this paper, we show how ranking can be used as a proxy task for some regression problems. As another contribution, we propose an efficient backpropagation technique for Siamese networks which prevents the redundant computation introduced by the multi-branch network architecture. We apply our framework to two regression problems: Image Quality Assessment (IQA) and Crowd Counting. For both we show how to automatically generate ranked image sets from unlabeled data. Our results show that networks trained to regress to the ground truth targets for labeled data and to simultaneously learn to rank unlabeled data obtain significantly better, state-of-the-art results for both IQA and crowd counting. In addition, we show that measuring network uncertainty on the self-supervised proxy task is a good measure of informativeness of unlabeled data. This can be used to drive an algorithm for active learning and we show that this reduces labeling effort by up to 50 percent.

Abstract: Optical Music Recognition (OMR) is the branch of document image analysis that aims to convert images of musical scores into a computer-readable format. Despite decades of research, the recognition of handwritten music scores, concretely the Western notation, is still an open problem, and the few existing works only focus on a specific stage of OMR. In this work, we propose a full Handwritten Music Recognition (HMR) system based on Convolutional Recurrent Neural Networks, data augmentation and transfer learning, that can serve as a baseline for the research community.

Abstract: Se analizan los sistemas de procesamiento automático que trabajan sobre documentos digitalizados con el objetivo de describir los contenidos. De esta forma contribuyen a facilitar el acceso, permitir la indización automática y hacer accesibles los documentos a los motores de búsqueda. El objetivo de estas tecnologías es poder entrenar modelos computacionales que sean capaces de clasificar, agrupar o realizar búsquedas sobre documentos digitales. Así, se describen las tareas de clasificación, agrupamiento y búsqueda. Cuando utilizamos tecnologías de inteligencia artificial en los sistemas de
clasificación esperamos que la herramienta nos devuelva etiquetas semánticas; en sistemas de agrupamiento que nos devuelva documentos agrupados en clusters significativos; y en sistemas de búsqueda esperamos que dada una consulta, nos devuelva una lista ordenada de documentos en función de la relevancia. A continuación se da una visión de conjunto de los métodos que nos permiten describir los documentos digitales, tanto de manera visual (cuál es su apariencia), como a partir de sus contenidos semánticos (de qué hablan). En cuanto a la descripción visual de documentos se aborda el estado de la cuestión de las representaciones numéricas de documentos digitalizados
tanto por métodos clásicos como por métodos basados en el aprendizaje profundo (deep learning). Respecto de la descripción semántica de los contenidos se analizan técnicas como el reconocimiento óptico de caracteres (OCR); el cálculo de estadísticas básicas sobre la aparición de las diferentes palabras en un texto (bag-of-words model); y los métodos basados en aprendizaje profundo como el método word2vec, basado en una red neuronal que, dadas unas cuantas palabras de un texto, debe predecir cuál será la
siguiente palabra. Desde el campo de las ingenierías se están transfiriendo conocimientos que se han integrado en productos o servicios en los ámbitos de la archivística, la biblioteconomía, la documentación y las plataformas de gran consumo, sin embargo los algoritmos deben ser lo suficientemente eficientes no sólo para el reconocimiento y transcripción literal sino también para la capacidad de interpretación de los contenidos.

Abstract: Food plays an important role in several aspects of our daily life. Several computer vision approaches have been proposed for tackling food analysis problems, but very little effort has been done in developing methodologies that could take profit of the existent correlation between tasks. In this paper, we propose a new multi-task model that is able to simultaneously predict different food-related tasks, e.g. dish, cuisine and food categories. Here, we extend the homoscedastic uncertainty modeling to allow single-label and multi-label classification and propose a regularization term, which jointly weighs the tasks as well as their correlations. Furthermore, we propose a new Multi-Attribute Food dataset and a new metric, Multi-Task Accuracy. We prove that using both our uncertainty-based loss and the class regularization term, we are able to improve the coherence of outputs between different tasks. Moreover, we outperform the use of task-specific models on classical measures like accuracy or .

Abstract: Blur detection aims at segmenting the blurred areas of a given image. Recent deep learning-based methods approach this problem by learning an end-to-end mapping between the blurred input and a binary mask representing the localization of its blurred areas. Nevertheless, the effectiveness of such deep models is limited due to the scarcity of datasets annotated in terms of blur segmentation, as blur annotation is labor intensive. In this work, we bypass the need for such annotated datasets for end-to-end learning, and instead rely on object proposals and a model for blur generation in order to produce a dataset of synthetically blurred images. This allows us to perform self-supervised learning over the generated image and ground truth blur mask pairs using CNNs, defining a framework that can be employed in purely self-supervised, weakly supervised or semi-supervised configurations. Interestingly, experimental results of such setups over the largest blur segmentation datasets available show that this approach achieves state of the art results in blur segmentation, even without ever observing any real blurred image.

Abstract: Deep video action recognition models have been highly successful in recent years but require large quantities of manually-annotated data, which are expensive and laborious to obtain. In this work, we investigate the generation of synthetic training data for video action recognition, as synthetic data have been successfully used to supervise models for a variety of other computer vision tasks. We propose an interpretable parametric generative model of human action videos that relies on procedural generation, physics models and other components of modern game engines. With this model we generate a diverse, realistic, and physically plausible dataset of human action videos, called PHAV for “Procedural Human Action Videos”. PHAV contains a total of 39,982 videos, with more than 1000 examples for each of 35 action categories. Our video generation approach is not limited to existing motion capture sequences: 14 of these 35 categories are procedurally-defined synthetic actions. In addition, each video is represented with 6 different data modalities, including RGB, optical flow and pixel-level semantic labels. These modalities are generated almost simultaneously using the Multiple Render Targets feature of modern GPUs. In order to leverage PHAV, we introduce a deep multi-task (i.e. that considers action classes from multiple datasets) representation learning architecture that is able to simultaneously learn from synthetic and real video datasets, even when their action categories differ. Our experiments on the UCF-101 and HMDB-51 benchmarks suggest that combining our large set of synthetic videos with small real-world datasets can boost recognition performance. Our approach also significantly outperforms video representations produced by fine-tuning state-of-the-art unsupervised generative models of videos.