Catálogo de publicaciones - revistas

<jats:p>Mobile devices such as smartphones and autonomous vehicles increasingly rely on deep neural networks (DNNs) to execute complex inference tasks such as image classification and speech recognition, among others. However, continuously executing the entire DNN on mobile devices can quickly deplete their battery. Although task offloading to cloud/edge servers may decrease the mobile device’s computational burden, erratic patterns in channel quality, network, and edge server load can lead to a significant delay in task execution. Recently, approaches based on split computing (SC) have been proposed, where the DNN is split into a head and a tail model, executed respectively on the mobile device and on the edge server. Ultimately, this may reduce bandwidth usage as well as energy consumption. Another approach, called early exiting (EE), trains models to embed multiple “exits” earlier in the architecture, each providing increasingly higher target accuracy. Therefore, the trade-off between accuracy and delay can be tuned according to the current conditions or application demands. In this paper, we provide a comprehensive survey of the state of the art in SC and EE strategies by presenting a comparison of the most relevant approaches. We conclude the paper by providing a set of compelling research challenges.</jats:p>

<jats:p>Object pose detection and tracking has recently attracted increasing attention due to its wide applications in many areas, such as autonomous driving, robotics, and augmented reality. Among methods for object pose detection and tracking, deep learning is the most promising one that has shown better performance than others. However, survey study about the latest development of deep learning-based methods is lacking. Therefore, this study presents a comprehensive review of recent progress in object pose detection and tracking that belongs to the deep learning technical route. To achieve a more thorough introduction, the scope of this study is limited to methods taking monocular RGB/RGBD data as input and covering three kinds of major tasks: instance-level monocular object pose detection, category-level monocular object pose detection, and monocular object pose tracking. In our work, metrics, datasets, and methods of both detection and tracking are presented in detail. Comparative results of current state-of-the-art methods on several publicly available datasets are also presented, together with insightful observations and inspiring future research directions.</jats:p>

<jats:p>A generative adversarial network (GAN) is one of the most significant research directions in the field of artificial intelligence, and its superior data generation capability has garnered wide attention. In this paper, we discuss the recent advancements in GANs, particularly in the medical field. First, the different medical imaging modalities and the principal theory of GANs were analyzed and summarized, after which, the evaluation metrics and training issues were determined. Third, the extension models of GANs were classified and introduced one by one. Fourth, the applications of GAN in medical images including cross-modality, augmentation, detection, classification, and reconstruction were illustrated. Finally, the problems we needed to resolve, and future directions were discussed. The objective of this review is to provide a comprehensive overview of the GAN, simplify the GAN’s basics, and present the most successful applications in different scenarios.</jats:p>

<jats:p>Interpretability, explainability and transparency are key issues to introducing Artificial Intelligence methods in many critical domains: This is important due to ethical concerns and trust issues strongly connected to reliability, robustness, auditability and fairness, and has important consequences towards keeping the human in the loop in high levels of automation, especially in critical cases for decision making, where both (human and the machine) play important roles. While the research community has given much attention to explainability of closed (or black) prediction boxes, there are tremendous needs for explainability of closed-box methods that support agents to act autonomously in the real world. Reinforcement learning methods, and especially their deep versions, are such closed-box methods. In this article we aim to provide a review of state of the art methods for explainable deep reinforcement learning methods, taking also into account the needs of human operators - i.e., of those that take the actual and critical decisions in solving real-world problems. We provide a formal specification of the deep reinforcement learning explainability problems, and we identify the necessary components of a general explainable reinforcement learning framework. Based on these, we provide a comprehensive review of state of the art methods, categorizing them in classes according to the paradigm they follow, the interpretable models they use, and the surface representation of explanations provided. The article concludes identifying open questions and important challenges.</jats:p>

<jats:p>Recently, Generative Adversarial Networks (GANs) have received enormous progress, which makes them able to learn complex data distributions in particular faces. More and more efficient GAN architectures have been designed and proposed to learn the different variations of faces, such as cross pose, age, expression and style. These GAN based approaches need to be reviewed, discussed, and categorized in terms of architectures, applications, and metrics. Several reviews that focus on the use and advances of GAN in general have been proposed. However, the GAN models applied to the face, that we call facial GANs, have never been addressed. In this article, we review facial GANs and their different applications. We mainly focus on architectures, problems and performance evaluation with respect to each application and used datasets. More precisely, we reviewed the progress of architectures and we discussed the contributions and limits of each. Then, we exposed the encountered problems of facial GANs and proposed solutions to handle them. Additionally, as GANs evaluation has become a notable current defiance, we investigate the state of the art quantitative and qualitative evaluation metrics and their applications. We concluded the article with a discussion on the face generation challenges and proposed open research issues.</jats:p>

<jats:p>Estimation of the human pose from a monocular camera has been an emerging research topic in the computer vision community with many applications. Recently, benefiting from the deep learning technologies, a significant amount of research efforts have advanced the monocular human pose estimation both in 2D and 3D areas. Although there have been some works to summarize different approaches, it still remains challenging for researchers to have an in-depth view of how these approaches work from 2D to 3D. In this paper, we provide a comprehensive and holistic 2D-to-3D perspective to tackle this problem. Firstly, we comprehensively summarize the 2D and 3D representations of human body. Then we summarize the mainstream and milestone approaches for these human body presentations since the year 2014 under unified frameworks. Especially, we provide insightful analyses for the intrinsic connections and methods evolution from 2D to 3D pose estimation. Furthermore, we analyze the solutions for challenging cases, such as the lack of data, the inherent ambiguity between 2D and 3D, and the complex multi-person scenarios. Next, we summarize the benchmarks, evaluation metrics, and the quantitative performance of popular approaches. Finally, we discuss the challenges and give deep thinking of promising directions for future research. We believe this survey will provide the readers (researchers, engineers, developers, etc.) with a deep and insightful understanding of monocular human pose estimation.</jats:p>

<jats:p>Two decades ago, a breakthrough in indexing string collections made it possible to represent them within their compressed space while at the same time offering indexed search functionalities. As this new technology permeated through applications like bioinformatics, the string collections experienced a growth that outperforms Moore’s Law and challenges our ability to handle them even in compressed form. It turns out, fortunately, that many of these rapidly growing string collections are highly repetitive, so that their information content is orders of magnitude lower than their plain size. The statistical compression methods used for classical collections, however, are blind to this repetitiveness, and therefore a new set of techniques has been developed to properly exploit it. The resulting indexes form a new generation of data structures able to handle the huge repetitive string collections that we are facing. In this survey, formed by two parts, we cover the algorithmic developments that have led to these data structures.</jats:p> <jats:p>In this first part, we describe the distinct compression paradigms that have been used to exploit repetitiveness, and the algorithmic techniques that provide direct access to the compressed strings. In the quest for an ideal measure of repetitiveness, we uncover a fascinating web of relations between those measures, as well as the limits up to which the data can be recovered, and up to which direct access to the compressed data can be provided. This is the basic aspect of indexability, which is covered in the second part of this survey.</jats:p>

<jats:p>Two decades ago, a breakthrough in indexing string collections made it possible to represent them within their compressed space while at the same time offering indexed search functionalities. As this new technology permeated through applications like bioinformatics, the string collections experienced a growth that outperforms Moore’s Law and challenges our ability of handling them even in compressed form. It turns out, fortunately, that many of these rapidly growing string collections are highly repetitive, so that their information content is orders of magnitude lower than their plain size. The statistical compression methods used for classical collections, however, are blind to this repetitiveness, and therefore a new set of techniques has been developed to properly exploit it. The resulting indexes form a new generation of data structures able to handle the huge repetitive string collections that we are facing. In this survey, formed by two parts, we cover the algorithmic developments that have led to these data structures.</jats:p> <jats:p>In this second part, we describe the fundamental algorithmic ideas and data structures that form the base of all the existing indexes, and the various concrete structures that have been proposed, comparing them both in theoretical and practical aspects, and uncovering some new combinations. We conclude with the current challenges in this fascinating field.</jats:p>

<jats:p>The history of text can be traced back over thousands of years. Rich and precise semantic information carried by text is important in a wide range of vision-based application scenarios. Therefore, text recognition in natural scenes has been an active research topic in computer vision and pattern recognition. In recent years, with the rise and development of deep learning, numerous methods have shown promising results in terms of innovation, practicality, and efficiency. This article aims to (1) summarize the fundamental problems and the state-of-the-art associated with scene text recognition, (2) introduce new insights and ideas, (3) provide a comprehensive review of publicly available resources, and (4) point out directions for future work. In summary, this literature review attempts to present an entire picture of the field of scene text recognition. It provides a comprehensive reference for people entering this field and could be helpful in inspiring future research. Related resources are available at our GitHub repository: https://github.com/HCIILAB/Scene-Text-Recognition.</jats:p>

<jats:p> With the development of the Internet of Things (IoT), many kinds of wireless signals (e.g., Wi-Fi, LoRa, RFID) are filling our living and working spaces nowadays. Beyond communication, wireless signals can sense the status of surrounding objects, known as <jats:italic>wireless sensing</jats:italic> , with their reflection, scattering, and refraction while propagating in space. In the last decade, many sophisticated wireless sensing techniques and systems were widely studied for various applications (e.g., gesture recognition, localization, and object imaging). Recently, deep Artificial Intelligence (AI), also known as Deep Learning (DL), has shown great success in computer vision. And some works have initially proved that deep AI can benefit wireless sensing as well, leading to a brand-new step toward ubiquitous sensing. In this survey, we focus on the evolution of wireless sensing enhanced by deep AI techniques. We first present a general workflow of Wireless Sensing Systems (WSSs) which consists of signal pre-processing, high-level feature, and sensing model formulation. For each module, existing deep AI-based techniques are summarized, further compared with traditional approaches. Then, we provide a view of issues and challenges induced by combining deep AI and wireless sensing together. Finally, we discuss the future trends of deep AI to enable ubiquitous wireless sensing. </jats:p>