Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)最新文献

{"title":"Segmentação Multi-classes em termografias mamárias utilizando Redes Profundas","authors":"Gabriela Pinheiro Henriger, S. Pinto","doi":"10.5753/sibgrapi.est.2022.23268","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23268","url":null,"abstract":"Apesar dos muitos avanços da medicina e da ciência no combate ao câncer de mama, estudos recentes sobre a incidência da doença no Brasil e no mundo mostram que essa doença é uma das principais causas de morte entre as mulheres. A fim de colaborar com o diagnóstico precoce das anomalias mamárias, aumentando as chances de cura, a imagem da termografia mamária tem sido utilizada com o intuito de colaborar onde a mamografia é desfavorável. Assim, este trabalho propõe a utilização do protocolo estático de termografia, pois oferece 5 pontos de vista da região mamária, ampliando assim a possibilidade de encontrar carcinomas. E, para encontrar a região de interesse nessas imagens, utilizamos a arquitetura de rede U-Net para realizar uma segmentação multi-classe com o objetivo de fornecer uma etapa futura de reconhecimento de padrões termográficos mais eficientes entre mamas saudáveis e anormais. Os resultados preliminares alcançaram uma precisão de 80,71% e um valor de 0,82% para a métrica IoU.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122173181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Visual crime pattern analysis","authors":"Germain García-Zanabria, L. G. Nonato","doi":"10.5753/sibgrapi.est.2022.23261","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23261","url":null,"abstract":"Studying and analyzing crime patterns in big cities is a challenging Spatio-temporal problem. The problem’s difficulty is linked to different factors such as data modeling, unsophisticated hotspot detection techniques, Spatio-temporal patterns, and study delimitation. Previous works have mostly focused on the analysis of crimes with the intent of uncovering patterns associated to social factors, seasonality, and urban activities in whole districts, regions, and neighborhoods. Those tools can hardly allow micro-scale crime analysis closely related to crime opportunity, whose understanding is fundamental for planning preventive actions. Visualizing different patterns hidden in crime time series data is another issue in this context, mainly due to the number of patterns that can show up in the time series analysis. In this dissertation, we propose a set of approaches for interactive visual crime analysis. Relying on machine learning methods, statistical and mathematical mechanisms, and visualization, each proposed methodology focus on solving specific crime-related problems. These proposed tools to explore specific city locations turned out to be essential for domain experts to accomplish their analysis in a bottom-up fashion, revealing how urban features related to mobility, passerby behavior, and the presence of public infrastructures can influence the quantity and type of crimes. The effectiveness and usefulness of the proposed methodologies have been demonstrated with a comprehensive set of quantitative and qualitative analyses, as well as case studies performed by domain experts involving real data from different-sized cities. The experiments show the capability of our approaches in identifying different crime-related phenomena.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131331064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Face-capture with automatic blendshape generation","authors":"João Vitor Nogueira, Bruno Sumar, Leonardo Carvalho","doi":"10.5753/sibgrapi.est.2022.23269","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23269","url":null,"abstract":"The goal of this paper is the development of a system for the production of performance-driven facial animation that automatically generates a blendshape model from input video frames. This simplifies the production of this kind of model to be used in animation projects. The proposed method is designed to be used without the need for expensive hardware, such that any computer with a webcam can run the system.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134070461","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Iterative Optimum-Path Forest: A Graph-Based Data Clustering Framework","authors":"David Aparco-Cardenas, Alexander X. Falcão, P. J. Rezende","doi":"10.5753/sibgrapi.est.2022.23259","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23259","url":null,"abstract":"Data clustering is widely recognized as a fundamental technique of paramount importance in pattern recognition and data mining. It is extensively used in many fields of the sciences, business and engineering, covering a broad spectrum of applications. Despite the large number of clustering methods, only a few of them take advantage of optimum connectivity among samples for more effective clustering. In this work, we aim to fill this gap by introducing a novel graph-based data clustering framework, called Iterative Optimum-Path Forest (IOPF), that exploits optimum connectivity for the design of improved clustering methods. The IOPF framework consists of four fundamental components: (i) sampling of a seed set S, (ii) partition of the graph induced from the dataset samples by an Optimum-Path Forest (OPF) rooted at S, (iii) recomputation of S based on the previous graph partition, and, after multiple iterations of the last two steps, (iv) selection of the forest with the lowest total cost across all iterations. IOPF can be regarded as a generalization of the Iterative Spanning Forest (ISF) framework for superpixel segmentation from the image domain to the feature space. Herein, we present four IOPF-based clustering solutions to illustrate distinct choices of its constituent components. These are thereafter employed to address three different problems, namely, unsupervised object segmentation, road network analysis and clustering of synthetic two-dimensional datasets, in order to assess their effectiveness under various graph topologies, and to ascertain their efficacy and robustness when compared to competitive baselines.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131858120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Crime prediction and prevention using police patrolling data: challenges and prospects","authors":"Thales Vieira, Tiago Paulino, João Matheus Siqueira Souza, Edival Lima","doi":"10.5753/sibgrapi.est.2022.23285","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23285","url":null,"abstract":"Spatiotemporal crime analysis and prediction aim at identifying criminal patterns in space and time. In previous work, crime prediction has been performed by identifying hotspots from data, which means areas of high criminal activity on the streets. By focusing efforts on such sites, police patrolling is expected to be more efficient, thus reducing criminal activity. However, not many studies focus on investigating how police patrolling affects crime, and whether it can be a predictor of crime activity. In this paper we discuss the main challenges of this problem, and describe some work in progress towards developing a robust methodology to represent, visually analyze, and build predictors for criminal activity, considering both criminal and police patrolling spatiotemporal data. As a case study, we use real datasets from the Military Police of the state of Alagoas, Brazil (PM-AL).","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"505 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122756426","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Interactive Image Segmentation: From Graph-based Algorithms to Feature-Space Annotation","authors":"Jordão Bragantini, A. Falcão","doi":"10.5753/sibgrapi.est.2022.23260","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23260","url":null,"abstract":"In recent years, machine learning algorithms that solve problems from a collection of examples (i.e. labeled data), have grown to be the predominant approach for solving computer vision and image processing tasks. These algorithms’ performance is highly correlated with the abundance of examples and their quality, especially methods based on neural networks, which are significantly data-hungry. Notably, image segmentation annotation requires extensive effort to produce high-quality labeling due to the fine-scale of the units (pixels) and resorts to interactive methodologies to provide user assistance. Therefore, improving interactive image segmentation methodologies with the goal of improving data labeling problems is of paramount importance to advance applications of computer vision methods. With this in mind, we investigated the existing literature on interactive image segmentation, contributing to it by introducing novel algorithms that perform the segmentation from markers, contours, and finally proposing a new paradigm for image annotation at scale.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125957294","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D Point-Cloud Quality Assessment Using Color and Geometry Texture Descriptors","authors":"Rafael Diniz, P. Freitas, Mylène C. Q. Farias","doi":"10.5753/sibgrapi.est.2022.23254","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23254","url":null,"abstract":"Since the mid-20th century, the use of digital formats for visual content allowed a great evolution in how society communicates. The Internet and digital broadcast systems introduced in the decade 90 to the wider public allowed an incredible expansion of multimedia consumption by the people, while the telecommunication networks and providers were pushed to their limits to address the growing multimedia content demand. Older electronic imaging systems, notably TV broadcasting systems, were designed after long subjective quality analysis for the definition of parameters like the number of lines of the video. But recent digital visual content services need faster and more affordable ways of evaluating the human perceived quality of the always-evolving multimedia systems. To address the need for automatic quality assessment, in the past decades many visual quality models based on algorithms that run on digital computers have been proposed. While the existing models are remarkably advanced for 2D digital imagery, a new set of immersive media is dawning, with different data structures, to which the 2D methods are not applicable, and need novel quality assessment metrics. These novel dawning immersive media formats provide a 3D visual representation of real objects and scenes. In this new visual format, objects can be captured, compressed, transmitted, and visualized in real-time not anymore as a flat 2D image, but as 3D content, allowing free-viewpoint selection by a consumer of such media. One of the most popular formats for immersive media is Point Cloud (PC), which is composed of points with 3 geometry coordinates plus color information, and sometimes, other information like reflectance and transparency. This work presents a research on the quality assessment of 3D PC based on novel color and geometric texture statistics. Considering that distortions to both color and geometry attributes of 3D visual content affect the perceived visual quality, it is proposed in this work to use both color-based and geometry-based texture descriptors for PC to obtain the visual degradation through their statistics. This work introduces 4 novels PC texture descriptors, 3 of them color-based, while 1 is geometry-based. Also, a new voxelization method is proposed, which converts points to voxels (volume elements), and improves the performance of the color-based texture descriptors. The performance of the proposed PC quality assessment method is among the best of the state-of-the-art PC quality assessment methods while being flexible and extensible to adapt to different types of distortions.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124628328","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Weaklier Supervised: Semi-automatic Scribble Generation Applied to Semantic Segmentation","authors":"João Pedro Klock Ferreira, João Paulo Lara Pinto, C. Castro","doi":"10.5753/sibgrapi.est.2022.23266","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23266","url":null,"abstract":"With many applications regarding semantic segmentation arising, along with the advent of the Deep Semantic Segmentation Networks, the need for large labeled datasets has also largely increased. But labeling thousands of images can be very expensive and time-consuming. Approaches such as weak and semi supervision try do deal with this problem, but the first cannot deal with large datasets and the latter is hard to deal with semantic segmentation. Therefore, in this work we propose a combination of both to create a novel pipeline of weak supervision, with focus in satellite imagery, capable of dealing with large datasets. We propose a pipeline to automatically generate scribbles in images, requiring that the user only label 10% of the images in a given dataset, while a classifier deal with the remaining images. Along with that, we also propose a simple semantic segmentation pipeline, that uses only images with scribbles to train a network. Results show that performance is lower, but similar to a fully supervised pipeline.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121929859","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Generative Adversarial Network approach for automatic inspection in automotive assembly lines","authors":"J. D. Mumbelli, G. A. Guarneri, Y. K. Lopes, Dalcimar Casanova, Marcelo Teixeira","doi":"10.5753/sibgrapi.est.2022.23262","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23262","url":null,"abstract":"In manufacturing systems, quality of inspection is a critical issue. This can be conducted by humans, or by employing Computer Vision Systems (CVS) which are trained upon representative datasets of images to detect classes of defects that may occur. The construction of such datasets strongly limits the use of CVS methods, as the variety of defects has combinatorial nature. Alternatively, instead of recognizing defects, a system can be trained to detect non-defective standards, becoming appropriate for some application profiles. In flexible automotive manufacturing, for example, parts are assembled within a reduced set of correct combinations, while the amount of possible incorrect assembling is enormous. In this paper, we show how a CVS can be extended with a Deep Learning-based approach that exploits a Generative Adversarial Network (GAN) to detect non-defective production, eliminating the need for constructing defect image datasets. The proposal is tested over the assembly line of Renault, in Brazil. Results show that our method returns better accuracy in inspection, compared with the current CVS solution, besides generalizing better to different components inspection without having to modify the method.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131129960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Data Lake and Analytics Platform with Application to COVID-19 Dynamic Analysis","authors":"F. Pereira, J. G. F. S. Costa, L. M. Gonçalves","doi":"10.5753/sibgrapi.est.2022.23283","DOIUrl":"https://doi.org/10.5753/sibgrapi.est.2022.23283","url":null,"abstract":"We propose a platform consisting of a data lake that has been implemented as a web-based service, to specifically solve the Covid-19 data production and processing problem. The main idea is that it can be used by data scientists working on COVID-19-related projects in order to access as much data as possible in one repository and be able not only to analyze that data but also to manage and contribute to new data. Through this platform, it has been possible to dynamically aggregate different data repositories related to the COVID-19 pandemic, in order to provide users, through a web interface, tools for use, transformations, and collaboration of data, as well as analysis and visualization tools integrated to geographic information systems.","PeriodicalId":182158,"journal":{"name":"Anais Estendidos do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122303868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}