Sergio Lima Netto

Federal University of Rio de Janeiro | UFRJ · Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia (COPPE)

The Aedes aegypti mosquito transmits several diseases, including dengue, zika, and chikungunya. To prevent these diseases, identifying and removing mosquito breeding sites is essential, but it is a time-consuming and labor-intensive task. To improve efficiency, computer vision and machine learning can be used to detect potential breeding grounds au...

Every year, the Aedes aegypti mosquito infects millions of people with diseases such as dengue, zika, chikungunya, and urban yellow fever. The main form to combat these diseases is to avoid mosquito reproduction by searching for and eliminating the potential mosquito breeding grounds. In this work, we introduce a comprehensive dataset of aerial vid...

Early detection of oil spills is paramount for quick response measures. The use of images acquired by UAVs (unmanned aerial vehicles), combined with computer vision techniques and deep-learning, presents automation opportunities to oil spill surveillance activities. However, proper training of models capable of detecting objects and region of inter...

The massive growth of audiences eager for sport content has substantially increased workers’ demand in this profitable segment. Highlight identification is vital for summarizing football matches. ecision support tools can significantly reduce the number of company employees required to tackle such a task, widely benefiting workforce resource alloca...

This work addresses the problem of extracting events from human-written daily drilling reports (DDRs) in an automated way. Two distinct approaches based on an expert system and artificial intelligence techniques are proposed: rule-based language processing (RBLP) and deep neural networks (DNN). The RBLP employs regular expressions that are manually...

Condition-based monitoring of power-generation systems is naturally becoming a standard approach in industry due to its inherent capability of fast fault detection, thus improving system efficiency and reducing operational costs. Most such systems employ expertise-reliant rule-based methods. This work proposes a different framework, in which machin...

In this chapter, we introduce the concept of Bayesian Neural Network and motivate the reader, presenting its gains over the classical neural networks. We scrutinize four of the most popular algorithms in the area: Bayes by Backprop, Probabilistic Backpropagation, Monte Carlo Dropout, Variational Adam. Each algorithm has its peculiarities and approa...

In this chapter, we introduce generative models. We focus specifically on the Variational Autoencoder (VAE) family, which uses the same set of tools introduced in Chap. 3, but with a stark objective in mind. Here, we are interested in modeling the process that generates the observed data. This empowers us to simulate new data, create world models,...

It is increasingly common for students and practitioners in the Machine Learning (ML) field to lack the adequate tools of statistics and statistical inference, especially with the popularity of Deep Learning (DL). Although it is impossible to teach such a vast subject in two or three dozen pages, this chapter aims to introduce the basic concepts of...

In this chapter, we introduce the building blocks of Model-Based Machine Learning (MBML). We explain what it is and discuss its main enabling techniques: Bayesian inference, graphical models, and, more recently, probabilistic programming. Frequently, models are complex enough so that exact inference is not possible, and one must resort to approxima...

Recent outstanding results of supervised object detection in competitions and challenges are often associated with specific metrics and datasets. The evaluation of such methods applied in different contexts have increased the demand for annotated datasets. Annotation tools represent the location and size of objects in distinct formats, leading to a...

This book provides a straightforward look at the concepts, algorithms and advantages of Bayesian Deep Learning and Deep Generative Models. Starting from the model-based approach to Machine Learning, the authors motivate Probabilistic Graphical Models and show how Bayesian inference naturally lends itself to this framework. The authors present detai...

This papers deals with the automatic detection and classification of faulty events during the practical operation of oil and gas wells and lines. The events considered here are part of the publicly available 3W database developed by Petrobras, the Brazilian oil holding. Seven fault classes are considered, with distinct dynamics and patterns, as wel...

Every year, the \textit{Aedes aegypti} mosquito infects thousands of people with diseases such as dengue, zika, chikungunya, and urban yellow fever. The main form to combat these diseases is to avoid the transmitter reproduction by searching and eliminating the potential mosquito breeding grounds. In this work, we introduce a comprehensive database...

This work explores and compares the plethora of metrics for the performance evaluation of object-detection algorithms. Average precision (AP), for instance, is a popular metric for evaluating the accuracy of object detectors by estimating the area under the curve (AUC) of the precision × recall relationship. Depending on the point interpolation use...

This work presents a novel approach to automatically detect the position of dark fringes in birefringence images obtained during the flow of polymers through slit dies. The determination of these positions is important for rheology, as it allows one to obtain the principal stress difference (PSD) profile along the flow centerline. The developed app...

In this work, we present a baseline end-to-end system based on deep learning for automatic speech recognition in Brazilian Portuguese. To build such a model, we employ a speech corpus containing 158 hours of annotated speech by assembling four individual datasets, three of them publicly available, and a text corpus containing 10.2 millions of sente...

This paper presents a special matrix factorization based on sparse representation that detects anomalies in video sequences generated with moving cameras. Such representation is made by associating the frames of the target video, that is a sequence to be tested for the presence of anomalies, with the frames of an anomaly-free reference video, which...

This paper deals with the statistical modeling of key features of power line communication (PLC) channels that are necessary for designing data communication systems that operate over theses channels. The key features are average channel attenuation, root mean squared delay spread, coherence bandwidth and coherence time. All these features were est...

This paper addresses the problem of abandoned object detection in a cluttered environment using a camera moving along a straight track. The developed system compares captured images to a previously recorded reference video, thus requiring proper temporal alignment and geometric registration between the two signals. A real-time constraint is imposed...

This paper proposes two modifications in a classification method for unbalancing fault severity analysis in rotating machines based on the unbalancing mass force. The unbalancing severity was categorized into three severity levels, namely High (H), Medium (M) and Low (L). The feature vectors used information from discrete-time Fourier transform (DF...

Resumo-Todos os anos, milhares de pessoas são afetadas por doenças como dengue, chikungunya, zika e febre amarela. Todas essas doenças possuem umúnicoum´umúnico vetor em comum, o mosquito Aedes aegypti, que se reproduz emáguaem´emágua limpa e parada, usualmente acumulada em recipientes como pneus, garrafas, caixas d'´ agua etc. O uso de ferramentas...

This paper deals with the challenging problem of visual anomaly detection in a cluttered environment using videos acquired with a moving camera. The anomalies considered are abandoned objects. A new method is proposed for comparing two videos (an anomaly-free reference video and a target one possibly with anomalies) by using convolutional neural ne...

This work aims to increase the understanding and bring the attention to the use of the so-called hybrid power line communication-wireless (PLC-wireless) channel, which is here defined as the equivalent channel that results from the concatenation of PLC and wireless channels, for data communication purposes. In this regard, we discuss about statisti...

This paper proposes a method to detect anomalies in videos acquired by a camera mounted on a moving inspection robot. The proposed method is based on a spatio-temporal composition (STC) method, where a dense sampling is used to break the video into small 3D volumes that are used to calculate the probability of the spatio-temporal arrangements. This...

In a previous approach to the development of a character-level end-to-end automatic speech recognition (ASR) system using deep learning, most of the network mistakes seemed easy to identify. This paper investigates the effects of two different post-processing schemes towards the automatic correction of such errors without resorting to any complex d...

This paper presents a family of algorithms based on sparse decompositions that detect anomalies in video sequences obtained from slow moving cameras. These algorithms start by computing the union of subspaces that best represents all the frames from a reference (anomaly free) video as a low-rank projection plus a sparse residue. Then, they perform...

Efficient anomaly detection in surveillance videos across diverse environments represents a major challenge in Computer Vision. This paper proposes a background subtraction approach based on the recent deep learning framework of residual neural networks that is capable of detecting multiple objects of different sizes by pixel-wise foreground segmen...

This paper focuses on the characterization of Brazilian in-home power line channels for data communication when a sounding-based approach is applied. Based on a measurement campaign carried out in seven Brazilian residences, a statistical characterization of frequency response magnitude, average channel gain, coherence bandwidth, root mean squared...

This paper presents an open-source character-based end-to-end speech recognition system for Brazilian Portuguese (PT-BR). The first step of the work was the development of a PT-BR dataset—an ensemble of 4 previous datasets (of which 3 publicly available). The model trained on this dataset is a bidirectional long short-term memory network using conn...

This work proposes an automatic fault classifier that uses similarity-based modeling (SBM) to identify faults on rotating machines. The similarity model can be used either as an auxiliary model to generate features for a classifier or as a standalone classifier. A new approach for training the model using a prototype-selection method is investigate...

Similarity-based modeling (SBM) is a technique whereby the normal operation of a system is modeled in order to detect faults by analyzing their similarity to the normal system states. First proposed around two decades ago, SBM has been successfully used for fault detection in varied systems. In spite of this success, there is not much study perform...

This work focuses on the characterization of indoor hybrid power line communication (PLC)-wireless channels in the frequency band between 1.7 and 100 MHz. These hybrid channels allow the simultaneous exploitation of the ubiquitous PLC channel and the mobility benefits offered by the wireless signals radiating from and being induced into power cables...

This article presents a new approach to the problem of aligning data obtained from sensors in similar trajectories in different instants. The signals are aligned using correlation measures obtained from generic sensor ensembles that are synchronized with the signal of interest. The focus signals are generated in a closed-loop trajectory, so it is n...

In the CB-DoA algorithm, the slowest operation is replaced by two faster operations.

Big data analytics, applied in the industry to leverage data collection, processing and analysis, can allow a better understanding of production system's abnormal behavior. This knowledge is essential for the adoption of a proactive maintenance approach instead of conventional time-based strategies, leading to a paradigm shift towards Condition-Bas...

Covariance-based DoA estimation (CB-DoA) algorithms represent lower computational complexity alternatives to the traditional ESPRIT approach. This paper investigates CB-DoA using Krylov-subspace techniques (including Arnoldi’s and Lanczos’ updates) with respect to the resulting computational cost and estimation error performance. The proposed modif...

This work describes a complete statistical modeling of the average channel gain in dB (ACGdB) and the root mean squared delay spread (RMS-DS) for power line communication (PLC) systems. The PLC channel features are estimated from 148,037 channel frequency responses measured in 7 typical different places in an urban area in Brazil. Two frequency ban...

This paper addresses the problem of reducing the reverberation effect from speech signals, which is known as dereverberation. The main idea is to modify a dereverberation algorithm based on ideal channel selection (ICS) from an algorithm with reference to a blind algorithm. The channel selection technique performs a comparison between clean and deg...

Frequency-estimation algorithms devised for complex sinusoids, including the maximum-likelihood (ML) approach, when operating on real sinusoidal signals, suffer from spectral interference due to the superposition of the aliasing components at negative and positive frequencies. This paper introduces a frequency estimation ML-like algorithm, based on...

Currently, there is a lack of standard methodology to characterize frequency responses of electric power grids for power line communication purpose. As a result, fair comparisons among measurement campaigns carried out in different parts of the world are missing. Aiming at to deal with this issue, this contribution discusses a complete sounding-bas...

Application of De-reverberation algorithms to speech enhancement

Safety and efficient operation are imperative factors to offshore production sites and a main concern to all Oil & Gas companies. A promising solution to improve both safety and efficiency is to increase the level of automation on the platforms by introducing intelligent robotic systems. Robots can execute a wide variety of tasks in offshore enviro...

This paper considers the problem of quantifying the reverberation perception on speech signals. We investigate several combinations of three distinct reverberation-related features (namely, the reverberation time (RT), room spectral variance (RSV), and direct-to-reverberant energy ratio (DRR)), which can be extracted directly from the associated ro...

This work aims at measuring, characterizing, and analyzing hybrid channels for mobile data communication with power line communication (PLC) technologies. These channels denote the concatenation of power line and wireless as only one channel. A measurement setup and OFDM-based technique for channel estimation were applied to carry out a measurement...

The paper addresses the problem of classifying mechanical faults in rotating machines. In this context, three operational classes are considered, namely: normal (where the machine has no fault), unbalance (where the machine load has its weight not equally distributed), and misalignment (where the rotor and machine axes are dislocated from its natur...

This paper describes an optimization strategy based on a perceptual assessment criterion for dereverberation algorithms. The complete procedure is applied to the adaptive inversefiltering (AIF) and spectral subtraction (SS) stages of a given dereverberation algorithm using the so-called QAreverb quality measure. Experimental results, using a 204-si...

This article presents a new algorithm for performing direction-of-arrival (DOA) estimation using manipulations on covariance matrices. The proposed algorithm combines a new formulation for data projection on real subspaces, together with beamspace decompositions, reducing the sizes of all data structures and computational complexity of the resultin...

This paper presents a study of the capacity of four speech signal features to assess speech perceptual quality and their use in a typical two-stage algorithm for reverberant speech enhancement. This algorithm is divided into two blocks: one that deals with the coloration effect, due to the early reflections, and the other for reducing the long-term...

An algorithm for blind estimation of reverberation time (RT) in speech signals is proposed. Analysis is restricted to the free-decaying regions of the signal, where the reverberation effect dominates, yielding a more accurate RT estimate at a reduced computational cost. A spectral decomposition is performed on the reverberant signal and partial RT...

This paper addresses the problem of quantifying the reverberation effect in speech signals. The perception of reverberation is assessed based on a new measure combining the characteristics of reverberation time, room spectral variance, and direct-to-reverberant energy ratio, which are estimated from the associated room impulse response (RIR). The p...

proposing a single matrix that stores the complete description of the filter in a very compact and functional format. The proposed matrix contains the structural information corresponding to the block diagram (BD) connections and, at the same time, it can be seen as a valid computational algorithm to implement the filter in the time domain. With th...

Prosody transplantation is a speech signal modification procedure usually used to voice transformation or to evaluate the quality of speech synthesizers. In practice, the pitch contour is mapped onto a common segmental content and the target signal is modified adjusting position and length of speech frames to achieve the desired pitch contour and t...

This article presents a low-complexity parametric algorithm for the estimation of the carrier frequency offset (CFO) in orthogonal frequency division multiplexing systems. The proposed algorithm is equivalent to the Unitary ESPRIT algorithm, originally devised for estimating the directionof-arrival of a wavefront, now applied to the CFO scenario. I...

This paper presents a direct-data (DD) counterpart to the covariance-based (CB) algorithm for direction-of-arrival (DOA) estimation. The proposed DD-DOA scheme provides reduced computational complexity as compared with other ESPRIT variations, filling in a theoretical gap not covered by previously presented schemes. A mean-squared error (MSE) analy...

This paper revisits the waveform paradigm for coding speech signals, using a multiscale recurrent-pattern matching approach. The so-called MMP (Multidimensional Multiscale Parser) algorithm uses a dictionary which is constantly updated with expansions, contractions, and concatenations of previously encoded segments. This provides a learning ability...

This new, fully-revised edition covers all the major topics of digital signal processing (DSP) design and analysis in a single, all-inclusive volume, interweaving theory with real-world examples and design trade-offs. Building on the success of the original, this edition includes new material on random signal processing, a new chapter on spectral e...

We propose an automatic engine for panoramic-take detection which relies on an algorithm based on phase correlation and boosting. The motion between two sequential video frames is first estimated through a phase correlation. Then, we are able to extract motion parameters and apply post-processing operations on these parameters in order to feed an A...

This chapter introduces the general concepts of adaptive filtering and its families of algorithms, and settles the basic notation used in the remaining of the book. Section presents the fundamentals concepts, highlighting several configurations, such as system identification, interference cancelation, channel equalization, and signal prediction, in...

This paper analyzes the ability of several measurements to quantify the reverberation effect in speech signals. We consider an intrusive scheme, in which the clean and reverberated signals are available, allowing one to estimate the corresponding room impulse response (RIR) signal. An artificial neural network (ANN) is trained for all features and...

This paper describes computationally efficient implementations for the ITU-T G.729 speech codec. Focus is given to the adaptive codebook search, more specifically in the open-loop stage, which first estimates the pitch period of the speech frame being coded. Different strategies are discussed to achieve an excellent compromise between computational...

Modern telepresence systems constitute a new challenge for quality assessment of multimedia signals. This paper focuses on the evaluation of the reverberation impairment for audioband speech signals. A review on the reverberation effect is presented, with emphasis given on the mathematical modeling of its components, including early reflections and...

This work presents a new version, with reduced computational complexity, of the covariance-based direction-of-arrival (CB-DoA) algorithm. The new algorithm incorporates the concept of beamspace projection before performing the DoA estimation. Such modification reduces the dimensions of the matrices employed by the elements pace CB-DoA, simplifying...

Modern telepresence systems constitute a new challenge for quality assessment of multimedia signals. This paper focuses on the evaluation of the reverberation impairment for audioband speech signals. A review on the reverberation effect is presented, with emphasis given on the mathematical modeling of its components, including early reflections and...

This paper constitutes an introduction to the field of quality evaluation of sound (speech and audio) signals. The need for such an assessment is inherent to modern communications: VoIP, mobile phone, or teleconference systems require meaningful measures of performance, which may ultimately assure good service or profitable business. A brief survey...