\n \n \n
\n
\n\n \n \n \n \n \n \n Embedding Deep Learning Models into Hypermedia Applications.\n \n \n \n \n\n\n \n Antonio José G. Busson; Álan Livio V. Guedes; Sérgio Colcher; Ruy Luiz Milidiú; and Edward Hermann Haeusler\n\n\n \n\n\n\n In Roesler, V.; Barrére, E.; and Willrich, R., editor(s),
Special Topics in Multimedia, IoT and Web Technologies, pages 91–111. Springer International Publishing, 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 16 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@incollection{antonio_jose_g_busson_embedding_2020,\n\tlocation = {Cham},\n\ttitle = {Embedding Deep Learning Models into Hypermedia Applications},\n\tisbn = {978-3-030-35102-1},\n\turl = {https://doi.org/10.1007/978-3-030-35102-1_4},\n\tabstract = {Deep learning research has allowed significant advances in several areas of multimedia, especially in tasks related to speech processing, hearing, and computational vision. Particularly, recent usage scenarios in hypermedia domain already use such deep learning tasks to build applications that are sensitive to its media content semantics. However, the development of such scenarios is usually done from scratch. In particular, current hypermedia standards such as {HTML} do not fully support such kind of development. To support such development, we propose that a hypermedia language should be extended to support: (1) describe learning using structured media datasets; (2) recognize content semantics of the media elements in presentation time; (3) use the recognized semantics elements as events in during the multimedia. To illustrate our approach, we extended the {NCL} language, and its model {NCM}, to support such features. {NCL} (Nested Context Language) is the declarative language for developing interactive applications for Brazilian Digital {TV} and an {ITU}-T Recommendation for {IPTV} services. As a result of the work, it is presented a usage scenario to highlight how the extended {NCL} supports the development of content-aware hypermedia presentations, attesting the expressiveness and applicability of the model.},\n\tpages = {91--111},\n\tbooktitle = {Special Topics in Multimedia, {IoT} and Web Technologies},\n\tpublisher = {Springer International Publishing},\n\tauthor = {{Antonio José G. Busson} and {Álan Livio V. Guedes} and {Sérgio Colcher} and {Ruy Luiz Milidiú} and {Edward Hermann Haeusler}},\n\teditor = {Roesler, Valter and Barrére, Eduardo and Willrich, Roberto},\n\turlyear = {2020},\n\tyear = {2020},\n\tlangid = {english},\n}\n\n
\n
\n\n\n
\n Deep learning research has allowed significant advances in several areas of multimedia, especially in tasks related to speech processing, hearing, and computational vision. Particularly, recent usage scenarios in hypermedia domain already use such deep learning tasks to build applications that are sensitive to its media content semantics. However, the development of such scenarios is usually done from scratch. In particular, current hypermedia standards such as HTML do not fully support such kind of development. To support such development, we propose that a hypermedia language should be extended to support: (1) describe learning using structured media datasets; (2) recognize content semantics of the media elements in presentation time; (3) use the recognized semantics elements as events in during the multimedia. To illustrate our approach, we extended the NCL language, and its model NCM, to support such features. NCL (Nested Context Language) is the declarative language for developing interactive applications for Brazilian Digital TV and an ITU-T Recommendation for IPTV services. As a result of the work, it is presented a usage scenario to highlight how the extended NCL supports the development of content-aware hypermedia presentations, attesting the expressiveness and applicability of the model.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n An Authoring Model for Interactive 360 videos.\n \n \n \n\n\n \n Mendes, P. R. C.; Guedes, A. L.; Moraes, D. d. S.; Azevedo, R. G. d. A.; and Colcher, S.\n\n\n \n\n\n\n In
Proceedings of ICME 2020 Workshop: Tools for Creating XR Media Experiences, 2020. \n
\n\n
\n\n
\n\n
\n\n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n\n\n
\n
@inproceedings{mendes_authoring_2020,\n\tlocation = {London, {UK}},\n\ttitle = {An Authoring Model for Interactive 360 videos},\n\tabstract = {The recent availability of consumer-level head-mounted displays and omnidirectional cameras has been driving an explosion of 360 video content. Transforming the original recorded 360 video content in meaningful interactive multimedia presentations that support viewers in tasks such as learning, entertainment, and telepresence, however, is nottrivial and requires new tools. Such tools must provide an easy-to-use authoring model for the integration of different media objects and active user interface elements. In this paper, based on real-world scenarios for interactive 360 videos, we gather a set of requirements and propose a declarative authoring model to support authors in the process of designing and creating 360-degree interactive video presentations. As a case study, we implement different applications showing the expressiveness and completeness of the model in the scope of the target scenarios.},\n\teventtitle = {{ICME} 2020 Workshop: Tools for Creating {XR} Media Experiences},\n\tbooktitle = {Proceedings of {ICME} 2020 Workshop: Tools for Creating {XR} Media Experiences},\n\tauthor = {Mendes, Paulo R. C. and Guedes, Alan Livio and Moraes, Daniel de Sousa and Azevedo, Roberto G. de A. and Colcher, Sérgio},\n\tyear = {2020},\n\tkeywords = {virtual reality, Media, user interfaces, 360-degree interactive video presentations, 360-degree video, active user interface elements, authoring, cameras, consumer-level head-mounted displays, declarative authoring model, helmet mounted displays, interactive multimedia presentations, interactive video, multimedia computing, Navigation, omnidirectional cameras, Omnidirectional video, Solid modeling, Three-dimensional displays, Timing, Two dimensional displays, Videos},\n}\n\n
\n
\n\n\n
\n The recent availability of consumer-level head-mounted displays and omnidirectional cameras has been driving an explosion of 360 video content. Transforming the original recorded 360 video content in meaningful interactive multimedia presentations that support viewers in tasks such as learning, entertainment, and telepresence, however, is nottrivial and requires new tools. Such tools must provide an easy-to-use authoring model for the integration of different media objects and active user interface elements. In this paper, based on real-world scenarios for interactive 360 videos, we gather a set of requirements and propose a declarative authoring model to support authors in the process of designing and creating 360-degree interactive video presentations. As a case study, we implement different applications showing the expressiveness and completeness of the model in the scope of the target scenarios.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n Video Quality Enhancement Using Deep Learning-Based Prediction Models for Quantized DCT Coefficients in MPEG I-frames.\n \n \n \n\n\n \n Busson, A. J.; Mendes, P. R.; Moraes, D. d. S; da Veiga, Á. M; Guedes, Á. L.; and Colcher, S.\n\n\n \n\n\n\n In
2020 IEEE International Symposium on Multimedia (ISM), 2020. \n
\n\n
\n\n
\n\n
\n\n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n\n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@inproceedings{busson_video_2020,\n\ttitle = {Video Quality Enhancement Using Deep Learning-Based Prediction Models for Quantized {DCT} Coefficients in {MPEG} I-frames},\n\tbooktitle = {2020 {IEEE} International Symposium on Multimedia ({ISM})},\n\tauthor = {Busson, Antonio {JG} and Mendes, Paulo {RC} and Moraes, Daniel de S and da Veiga, Álvaro M and Guedes, Álan {LV} and Colcher, Sérgio},\n\tyear = {2020},\n}\n\n
\n
\n\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Decoder-Side Quality Enhancement of JPEG Images Using Deep Learning-Based Prediction Models for Quantized DCT Coefficients.\n \n \n \n \n\n\n \n Busson, A. J. G.; Mendes, P. R. C.; de S. Moraes, D.; da Veiga, Á. M. G.; Colcher, S.; and Guedes, Á. L. V.\n\n\n \n\n\n\n In
Proceedings of the Brazilian Symposium on Multimedia and the Web, of
WebMedia '20, pages 129–136, 2020. Association for Computing Machinery\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 9 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n \n \n \n \n \n \n\n\n\n
\n
@inproceedings{busson_decoder-side_2020,\n\tlocation = {New York, {NY}, {USA}},\n\ttitle = {Decoder-Side Quality Enhancement of {JPEG} Images Using Deep Learning-Based Prediction Models for Quantized {DCT} Coefficients},\n\tisbn = {978-1-4503-8196-3},\n\turl = {https://doi.org/10.1145/3428658.3430966},\n\tdoi = {10.1145/3428658.3430966},\n\tseries = {{WebMedia} '20},\n\tabstract = {Many recent works have successfully applied some types of Convolutional Neural Networks ({CNNs}) to reduce the noticeable distortion resulting from the lossy {JPEG} compression technique. Most of them are built upon the processing made on the spatial domain. In this work, we propose a {JPEG} image decoder that is purely based on the frequency-to-frequency domain: it reads the quantized {DCT} coefficients received from a low-quality image bitstream and, using a deep learning-based model, predicts the missing coefficients in order to recompose the same image with enhanced quality. In experiments with two datasets, our best model was able to improve from images with quantized {DCT} coefficients corresponding to a Qualityz Factor ({QF}) of 10 to enhanced quality images with {QF} slightly higher than 20.},\n\tpages = {129--136},\n\tbooktitle = {Proceedings of the Brazilian Symposium on Multimedia and the Web},\n\tpublisher = {Association for Computing Machinery},\n\tauthor = {Busson, Antonio José G. and Mendes, Paulo Renato C. and de S. Moraes, Daniel and da Veiga, Álvaro Mário G. and Colcher, Sérgio and Guedes, Álan Lívio V.},\n\turlyear = {2020},\n\tyear = {2020},\n\tkeywords = {Convolutional Neural Networks, {DCT}, Deep learning, Image compression, {JPEG}},\n}\n\n
\n
\n\n\n
\n Many recent works have successfully applied some types of Convolutional Neural Networks (CNNs) to reduce the noticeable distortion resulting from the lossy JPEG compression technique. Most of them are built upon the processing made on the spatial domain. In this work, we propose a JPEG image decoder that is purely based on the frequency-to-frequency domain: it reads the quantized DCT coefficients received from a low-quality image bitstream and, using a deep learning-based model, predicts the missing coefficients in order to recompose the same image with enhanced quality. In experiments with two datasets, our best model was able to improve from images with quantized DCT coefficients corresponding to a Qualityz Factor (QF) of 10 to enhanced quality images with QF slightly higher than 20.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n ListeningTV: Accessible Video using Interactive Audio Descriptions.\n \n \n \n \n\n\n \n Vieira, A. d. S.; Guedes, Á. L. V.; Moraes, D. d. S.; Madeira, L. R.; Colcher, S.; and Neto, C. d. S. S.\n\n\n \n\n\n\n ,71–74. 2020.\n
Conference Name: Anais Estendidos do XXVI Simpósio Brasileiro de Sistemas Multimídia e Web Publisher: SBC\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 8 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{vieira_listeningtv_2020,\n\ttitle = {{ListeningTV}: Accessible Video using Interactive Audio Descriptions},\n\trights = {Copyright (c)},\n\tissn = {2596-1683},\n\turl = {https://sol.sbc.org.br/index.php/webmedia_estendido/article/view/13065},\n\tdoi = {10.5753/webmedia_estendido.2020.13065},\n\tshorttitle = {{ListeningTV}},\n\tabstract = {Resumo\n\t\t\t\t\tPeople with visual impairments suffer from the incapacity to understand contextual information in videos, such as the place where characters are, or any other non-spoken actions in general. Some content creators address this issue by providing a secondary audio to describe such information, called Audio Descriptions ({ADs}). How- ever, some works in the literature have highlighted that people with visual impairment are usually not able to completely understand scene changes based only on characters’ voices or traditional {ADs}. Moreover, traditional {ADs} do not completely describe some of the important visual information, such as the background scenery (e.g. colors, furniture) and characters’ details (e.g. blond woman using a red dress). In this work, we propose incrementing the traditional {AD} techniques with the usage of interactive video features present in {TV} systems. More precisely, the proposed interactivity enables users to access specialized {AD} for different visual information (e.g., scene, scenario, character). To support the development of such interactive content, we present an application template, which helps to create the final interactive-enhanced video application. Asa proof of concept for our approach, we created an interactive {AD} for an independent video mainly composed of visual information, with only a few talks.},\n\tpages = {71--74},\n\tjournaltitle = {Anais Estendidos do Simpósio Brasileiro de Sistemas Multimídia e Web ({WebMedia})},\n\tauthor = {Vieira, Alex de Souza and Guedes, Álan Lívio V. and Moraes, Daniel de Sousa and Madeira, Lucas Ribeiro and Colcher, Sérgio and Neto, Carlos de S. Soares},\n\turlyear = {2020},\n\tyear = {2020},\n\tlangid = {portuguese},\n\tnote = {Conference Name: Anais Estendidos do {XXVI} Simpósio Brasileiro de Sistemas Multimídia e Web\nPublisher: {SBC}},\n}\n\n
\n
\n\n\n
\n Resumo People with visual impairments suffer from the incapacity to understand contextual information in videos, such as the place where characters are, or any other non-spoken actions in general. Some content creators address this issue by providing a secondary audio to describe such information, called Audio Descriptions (ADs). How- ever, some works in the literature have highlighted that people with visual impairment are usually not able to completely understand scene changes based only on characters’ voices or traditional ADs. Moreover, traditional ADs do not completely describe some of the important visual information, such as the background scenery (e.g. colors, furniture) and characters’ details (e.g. blond woman using a red dress). In this work, we propose incrementing the traditional AD techniques with the usage of interactive video features present in TV systems. More precisely, the proposed interactivity enables users to access specialized AD for different visual information (e.g., scene, scenario, character). To support the development of such interactive content, we present an application template, which helps to create the final interactive-enhanced video application. Asa proof of concept for our approach, we created an interactive AD for an independent video mainly composed of visual information, with only a few talks.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Shaping the Video Conferences of Tomorrow With AI.\n \n \n \n \n\n\n \n Mendes, P. R. C.; Vieira, E. S.; Freitas, P. V. A. d.; Busson, A. J. G.; Guedes, Á. L. V.; Neto, C. d. S. S.; and Colcher, S.\n\n\n \n\n\n\n ,165–168. 2020.\n
Conference Name: Anais Estendidos do XXVI Simpósio Brasileiro de Sistemas Multimídia e Web Publisher: SBC\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 8 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{mendes_shaping_2020,\n\ttitle = {Shaping the Video Conferences of Tomorrow With {AI}},\n\trights = {Copyright (c)},\n\tissn = {2596-1683},\n\turl = {https://sol.sbc.org.br/index.php/webmedia_estendido/article/view/13082},\n\tdoi = {10.5753/webmedia_estendido.2020.13082},\n\tabstract = {Resumo\n\t\t\t\t\tBefore the {COVID}-19 pandemic, video was already one of the main media used on the internet. During the pandemic, video conferencing services became even more important, coming to be one of the main instruments to enable most social and professional human activities. Given the social distancing policies, people are spending more time using these online services for working, learning, and also for leisure activities. Videoconferencing software became the standard communication for home-office and remote learning. Nevertheless, there are still a lot of issues to be addressed on these platforms, and many different aspects to be reexamined or investigated, such as ethical and user-experience issues, just to name a few. We argue that many of the current state-of-the-art techniques of Artificial Intelligence ({AI}) may help on enhancing video collabo- ration services, particularly the methods based on Deep Learning such as face and sentiment analyses, and video classification. In this paper, we present a future vision about how {AI} techniques may contribute to this upcoming videoconferencing-age.},\n\tpages = {165--168},\n\tjournaltitle = {Anais Estendidos do Simpósio Brasileiro de Sistemas Multimídia e Web ({WebMedia})},\n\tauthor = {Mendes, Paulo Renato C. and Vieira, Eduardo S. and Freitas, Pedro Vinicius A. de and Busson, Antonio José G. and Guedes, Álan Lívio V. and Neto, Carlos de Salles Soares and Colcher, Sérgio},\n\turlyear = {2020},\n\tyear = {2020},\n\tlangid = {portuguese},\n\tnote = {Conference Name: Anais Estendidos do {XXVI} Simpósio Brasileiro de Sistemas Multimídia e Web\nPublisher: {SBC}},\n}\n\n
\n
\n\n\n
\n Resumo Before the COVID-19 pandemic, video was already one of the main media used on the internet. During the pandemic, video conferencing services became even more important, coming to be one of the main instruments to enable most social and professional human activities. Given the social distancing policies, people are spending more time using these online services for working, learning, and also for leisure activities. Videoconferencing software became the standard communication for home-office and remote learning. Nevertheless, there are still a lot of issues to be addressed on these platforms, and many different aspects to be reexamined or investigated, such as ethical and user-experience issues, just to name a few. We argue that many of the current state-of-the-art techniques of Artificial Intelligence (AI) may help on enhancing video collabo- ration services, particularly the methods based on Deep Learning such as face and sentiment analyses, and video classification. In this paper, we present a future vision about how AI techniques may contribute to this upcoming videoconferencing-age.\n
\n\n\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Relacionando Modelagem de Tópicos e Classificação de Sentimentos para Análise de Mensagens do Twitter Durante a Pandemia da COVID-19.\n \n \n \n \n\n\n \n Pinto, M. A. S.; Junior, A. F. L. J.; Busson, A. J. G.; and Colcher, S.\n\n\n \n\n\n\n ,61–64. 2020.\n
Conference Name: Anais Estendidos do XXVI Simpósio Brasileiro de Sistemas Multimídia e Web Publisher: SBC\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 8 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{pinto_relacionando_2020,\n\ttitle = {Relacionando Modelagem de Tópicos e Classificação de Sentimentos para Análise de Mensagens do Twitter Durante a Pandemia da {COVID}-19},\n\trights = {Copyright (c)},\n\tissn = {2596-1683},\n\turl = {https://sol.sbc.org.br/index.php/webmedia_estendido/article/view/13064},\n\tdoi = {10.5753/webmedia_estendido.2020.13064},\n\tabstract = {Resumo\n\t\t\t\t\tIn 2020, {COVID}-19 pandemic is one of the most talked-about subjects on social networks. This subject has generated discussions of great importance about politics, economics, medical advances, people’s awareness, preventive techniques, etc. Using sentiment analysis and topic modeling techniques, in this paper, we aim to present an analysis of the messages from the social network Twitter during the pandemic of {COVID}-19. For this, we use a tweets dataset to train a sentiment classifier and then use the {NMF} algorithm to perform the interest topic generation.},\n\tpages = {61--64},\n\tjournaltitle = {Anais Estendidos do Simpósio Brasileiro de Sistemas Multimídia e Web ({WebMedia})},\n\tauthor = {Pinto, Matheus Adler Soares and Junior, Antonio Fernando Lavareda Jacob and Busson, Antonio José G. and Colcher, Sérgio},\n\turlyear = {2020},\n\tyear = {2020},\n\tlangid = {portuguese},\n\tnote = {Conference Name: Anais Estendidos do {XXVI} Simpósio Brasileiro de Sistemas Multimídia e Web\nPublisher: {SBC}},\n}\n\n
\n
\n\n\n
\n Resumo In 2020, COVID-19 pandemic is one of the most talked-about subjects on social networks. This subject has generated discussions of great importance about politics, economics, medical advances, people’s awareness, preventive techniques, etc. Using sentiment analysis and topic modeling techniques, in this paper, we aim to present an analysis of the messages from the social network Twitter during the pandemic of COVID-19. For this, we use a tweets dataset to train a sentiment classifier and then use the NMF algorithm to perform the interest topic generation.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n ÁGATA: um chatbot para difusão de práticas para Educação Ambiental.\n \n \n \n \n\n\n \n Gomes, B. R.; Jr, A. F. L. J.; Pinto, I. d. J. P.; and Colcher, S.\n\n\n \n\n\n\n ,85–89. 2020.\n
Conference Name: Anais Estendidos do XXVI Simpósio Brasileiro de Sistemas Multimídia e Web Publisher: SBC\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 8 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{gomes_agata_2020,\n\ttitle = {Á{GATA}: um chatbot para difusão de práticas para Educação Ambiental},\n\trights = {Copyright (c)},\n\tissn = {2596-1683},\n\turl = {https://sol.sbc.org.br/index.php/webmedia_estendido/article/view/13068},\n\tdoi = {10.5753/webmedia_estendido.2020.13068},\n\tshorttitle = {Á{GATA}},\n\tabstract = {Resumo\n\t\t\t\t\tInstant messaging apps have billions of monthly active users, which made this environment a good medium for purposes other than communication, such as education. In this context, one of the most appropriate technologies to encourage learning in these applications is chatbots. They consist of conversational applications that make use of artificial intelligence and natural language processing to interact with the user in a similar way to a human. In this scenario, this article presents the {AGATE} a chatbot developed for the Environmental Management Office ({AGA} / {UEMA}) in order to disseminate knowledge on environmental education, focusing on the problems of water and energy waste. The chatbot was developed for Telegram through the {pyTelegramBotAPI} implementation. In addition, conversation flows were created using {DialogFlow}. A multidisciplinary team composed of 14 people participated in the evaluation of the project, with a greater participation of professionals in the field of Biology. The initial results focused on the user experience were satisfactory.},\n\tpages = {85--89},\n\tjournaltitle = {Anais Estendidos do Simpósio Brasileiro de Sistemas Multimídia e Web ({WebMedia})},\n\tauthor = {Gomes, Bruno Rocha and Jr, Antonio Fernando Lavareda Jacob and Pinto, Ivan de Jesus Pereira and Colcher, Sérgio},\n\turlyear = {2020},\n\tyear = {2020},\n\tlangid = {portuguese},\n\tnote = {Conference Name: Anais Estendidos do {XXVI} Simpósio Brasileiro de Sistemas Multimídia e Web\nPublisher: {SBC}},\n}\n\n
\n
\n\n\n
\n Resumo Instant messaging apps have billions of monthly active users, which made this environment a good medium for purposes other than communication, such as education. In this context, one of the most appropriate technologies to encourage learning in these applications is chatbots. They consist of conversational applications that make use of artificial intelligence and natural language processing to interact with the user in a similar way to a human. In this scenario, this article presents the AGATE a chatbot developed for the Environmental Management Office (AGA / UEMA) in order to disseminate knowledge on environmental education, focusing on the problems of water and energy waste. The chatbot was developed for Telegram through the pyTelegramBotAPI implementation. In addition, conversation flows were created using DialogFlow. A multidisciplinary team composed of 14 people participated in the evaluation of the project, with a greater participation of professionals in the field of Biology. The initial results focused on the user experience were satisfactory.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Prior Flow Variational Autoencoder: A density estimation model for Non-Intrusive Load Monitoring.\n \n \n \n \n\n\n \n Henriques, L. F. M. O.; Morgan, E.; Colcher, S.; and Milidiú, R. L.\n\n\n \n\n\n\n . 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 9 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n \n \n\n\n\n
\n
@article{henriques_prior_2020,\n\ttitle = {Prior Flow Variational Autoencoder: A density estimation model for Non-Intrusive Load Monitoring},\n\turl = {http://arxiv.org/abs/2011.14870},\n\tshorttitle = {Prior Flow Variational Autoencoder},\n\tabstract = {Non-Intrusive Load Monitoring ({NILM}) is a computational technique to estimate the power loads' appliance-by-appliance from the whole consumption measured by a single meter. In this paper, we propose a conditional density estimation model, based on deep neural networks, that joins a Conditional Variational Autoencoder with a Conditional Invertible Normalizing Flow model to estimate the individual appliance's power demand. The resulting model is called Prior Flow Variational Autoencoder or, for simplicity {PFVAE}. Thus, instead of having one model per appliance, the resulting model is responsible for estimating the power demand, appliance-by-appliance, at once. We train and evaluate our proposed model in a publicly available dataset composed of power demand measures from a poultry feed factory located in Brazil. The proposed model's quality is evaluated by comparing the obtained normalized disaggregation error ({NDE}) and signal aggregated error ({SAE}) with the previous work values on the same dataset. Our proposal achieves highly competitive results, and for six of the eight machines belonging to the dataset, we observe consistent improvements that go from 28\\% up to 81\\% in {NDE} and from 27\\% up to 86\\% in {SAE}.},\n\tjournaltitle = {{arXiv}:2011.14870 [cs]},\n\tauthor = {Henriques, Luis Felipe M. O. and Morgan, Eduardo and Colcher, Sergio and Milidiú, Ruy Luiz},\n\turlyear = {2021},\n\tyear = {2020},\n\teprinttype = {arxiv},\n\teprint = {2011.14870},\n\tkeywords = {Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning},\n}\n\n
\n
\n\n\n
\n Non-Intrusive Load Monitoring (NILM) is a computational technique to estimate the power loads' appliance-by-appliance from the whole consumption measured by a single meter. In this paper, we propose a conditional density estimation model, based on deep neural networks, that joins a Conditional Variational Autoencoder with a Conditional Invertible Normalizing Flow model to estimate the individual appliance's power demand. The resulting model is called Prior Flow Variational Autoencoder or, for simplicity PFVAE. Thus, instead of having one model per appliance, the resulting model is responsible for estimating the power demand, appliance-by-appliance, at once. We train and evaluate our proposed model in a publicly available dataset composed of power demand measures from a poultry feed factory located in Brazil. The proposed model's quality is evaluated by comparing the obtained normalized disaggregation error (NDE) and signal aggregated error (SAE) with the previous work values on the same dataset. Our proposal achieves highly competitive results, and for six of the eight machines belonging to the dataset, we observe consistent improvements that go from 28% up to 81% in NDE and from 27% up to 86% in SAE.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n Stocks Clustering Based on Textual Embeddings for Price Forecasting.\n \n \n \n\n\n \n de Oliveira, A. D. C. M.; Pinto, P. F. A.; and Colcher, S.\n\n\n \n\n\n\n In Cerri, R.; and Prati, R. C., editor(s),
Intelligent Systems, of
Lecture Notes in Computer Science, pages 665–678, 2020. Springer International Publishing\n
\n\n
\n\n
\n\n
\n\n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n \n \n\n\n\n
\n
@inproceedings{de_oliveira_stocks_2020,\n\tlocation = {Cham},\n\ttitle = {Stocks Clustering Based on Textual Embeddings for Price Forecasting},\n\tisbn = {978-3-030-61380-8},\n\tdoi = {10.1007/978-3-030-61380-8_45},\n\tseries = {Lecture Notes in Computer Science},\n\tabstract = {Forecasting stock market prices is a hard task. The main reason for that is due to the fact that its environment is highly dynamic, intrinsically complex, and chaotic. Traditional economic theories suggest that trying to forecast short-term stock price movements is a wasted effort because the market is influenced by several external events and its behavior approximates a random walk. Recent studies that address the problem of stock market forecasting usually create specific prediction models for the price behavior of a single stock. In this work we propose a technique to predict price movements based on similar stock sets. Our goal is to build a model to identify whether the price tends to bullishness or bearishness in the near future, considering stock information from similar sets based on two sources of information: historical stock data and Google Trends news. Firstly, the proposed study applies a method to identify similar stock sets and then creates a forecasting model based on a {LSTM} (long short-term memory) for these sets. More specifically, two experiments were conducted: (1) using the K-Means algorithm to identify similar stock sets and then using a {LSTM} neural network to forecast stock price movements for these stock sets; (2) using the {DBSCAN} (Density-based spatial clustering) algorithm to identify similar stock sets and then using the same {LSTM} neural network to forecast stock price movements. The study was conducted over 51 stocks of the Brazilian stock market. The results show that the use of an algorithm to identify stock clusters yields an improvement of approximately 7\\% in accuracy and f1-score and 8\\% in recall and precision when compared to models for a single stock.},\n\tpages = {665--678},\n\tbooktitle = {Intelligent Systems},\n\tpublisher = {Springer International Publishing},\n\tauthor = {de Oliveira, André D. C. M. and Pinto, Pedro F. A. and Colcher, Sergio},\n\teditor = {Cerri, Ricardo and Prati, Ronaldo C.},\n\tyear = {2020},\n\tlangid = {english},\n\tkeywords = {Forecasting time series, Machine learning, Stock market},\n}\n\n
\n
\n\n\n
\n Forecasting stock market prices is a hard task. The main reason for that is due to the fact that its environment is highly dynamic, intrinsically complex, and chaotic. Traditional economic theories suggest that trying to forecast short-term stock price movements is a wasted effort because the market is influenced by several external events and its behavior approximates a random walk. Recent studies that address the problem of stock market forecasting usually create specific prediction models for the price behavior of a single stock. In this work we propose a technique to predict price movements based on similar stock sets. Our goal is to build a model to identify whether the price tends to bullishness or bearishness in the near future, considering stock information from similar sets based on two sources of information: historical stock data and Google Trends news. Firstly, the proposed study applies a method to identify similar stock sets and then creates a forecasting model based on a LSTM (long short-term memory) for these sets. More specifically, two experiments were conducted: (1) using the K-Means algorithm to identify similar stock sets and then using a LSTM neural network to forecast stock price movements for these stock sets; (2) using the DBSCAN (Density-based spatial clustering) algorithm to identify similar stock sets and then using the same LSTM neural network to forecast stock price movements. The study was conducted over 51 stocks of the Brazilian stock market. The results show that the use of an algorithm to identify stock clusters yields an improvement of approximately 7% in accuracy and f1-score and 8% in recall and precision when compared to models for a single stock.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Seismic Shot Gather Noise Localization Using a Multi-Scale Feature-Fusion-Based Neural Network.\n \n \n \n \n\n\n \n Busson, A. J. G.; Colcher, S.; Milidiú, R. L.; Dias, B. P.; and Bulcão, A.\n\n\n \n\n\n\n . 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 6 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n\n\n\n
\n
@article{busson_seismic_2020,\n\ttitle = {Seismic Shot Gather Noise Localization Using a Multi-Scale Feature-Fusion-Based Neural Network},\n\turl = {http://arxiv.org/abs/2005.03626},\n\tabstract = {Deep learning-based models, such as convolutional neural networks, have advanced various segments of computer vision. However, this technology is rarely applied to seismic shot gather noise localization problem. This letter presents an investigation on the effectiveness of a multi-scale feature-fusion-based network for seismic shot-gather noise localization. Herein, we describe the following: (1) the construction of a real-world dataset of seismic noise localization based on 6,500 seismograms; (2) a multi-scale feature-fusion-based detector that uses the {MobileNet} combined with the Feature Pyramid Net as the backbone; and (3) the Single Shot multi-box detector for box classification/regression. Additionally, we propose the use of the Focal Loss function that improves the detector's prediction accuracy. The proposed detector achieves an {AP}@0.5 of 78.67{\\textbackslash}\\% in our empirical evaluation.},\n\tjournaltitle = {{arXiv}:2005.03626 [cs]},\n\tauthor = {Busson, Antonio José G. and Colcher, Sérgio and Milidiú, Ruy Luiz and Dias, Bruno Pereira and Bulcão, André},\n\turlyear = {2021},\n\tyear = {2020},\n\teprinttype = {arxiv},\n\teprint = {2005.03626},\n\tkeywords = {Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning},\n}\n\n
\n
\n\n\n
\n Deep learning-based models, such as convolutional neural networks, have advanced various segments of computer vision. However, this technology is rarely applied to seismic shot gather noise localization problem. This letter presents an investigation on the effectiveness of a multi-scale feature-fusion-based network for seismic shot-gather noise localization. Herein, we describe the following: (1) the construction of a real-world dataset of seismic noise localization based on 6,500 seismograms; (2) a multi-scale feature-fusion-based detector that uses the MobileNet combined with the Feature Pyramid Net as the backbone; and (3) the Single Shot multi-box detector for box classification/regression. Additionally, we propose the use of the Focal Loss function that improves the detector's prediction accuracy. The proposed detector achieves an AP@0.5 of 78.67\\% in our empirical evaluation.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n A Deep Learning Approach to Detect Pornography Videos in Educational Repositories.\n \n \n \n \n\n\n \n Freitas, P. V. A. d.; Busson, A. J. G.; Guedes, Á. L. V.; and Colcher, S.\n\n\n \n\n\n\n In
Anais do Simpósio Brasileiro de Informática na Educação, pages 1253–1262, 2020. \n
Conference Name: Anais do XXXI Simpósio Brasileiro de Informática na Educação Publisher: SBC\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 6 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@inproceedings{freitas_deep_2020,\n\ttitle = {A Deep Learning Approach to Detect Pornography Videos in Educational Repositories},\n\trights = {Copyright (c)},\n\turl = {https://sol.sbc.org.br/index.php/sbie/article/view/12881},\n\tdoi = {10.5753/cbie.sbie.2020.1253},\n\tabstract = {Resumo\n\t\t\t\t\tA large number of videos are uploaded on educational platforms every minute. Those platforms are responsible for any sensitive media uploaded by their users. An automated detection system to identify pornographic content could assist human workers by pre-selecting suspicious videos. In this paper, we propose a multimodal approach to adult content detection. We use two Deep Convolutional Neural Networks to extract high-level features from both image and audio sources of a video. Then, we concatenate those features and evaluate the performance of classifiers on a set of mixed educational and pornographic videos. We achieve an F1-score of 95.67\\% on the educational and adult videos set and an F1-score of 94\\% on our test subset for the pornographic class.},\n\tpages = {1253--1262},\n\tbooktitle = {Anais do Simpósio Brasileiro de Informática na Educação},\n\tauthor = {Freitas, Pedro V. A. de and Busson, Antonio J. G. and Guedes, Álan L. V. and Colcher, Sérgio},\n\turlyear = {2021},\n\tyear = {2020},\n\tlangid = {english},\n\tnote = {Conference Name: Anais do {XXXI} Simpósio Brasileiro de Informática na Educação\nPublisher: {SBC}},\n}\n\n
\n
\n\n\n
\n Resumo A large number of videos are uploaded on educational platforms every minute. Those platforms are responsible for any sensitive media uploaded by their users. An automated detection system to identify pornographic content could assist human workers by pre-selecting suspicious videos. In this paper, we propose a multimodal approach to adult content detection. We use two Deep Convolutional Neural Networks to extract high-level features from both image and audio sources of a video. Then, we concatenate those features and evaluate the performance of classifiers on a set of mixed educational and pornographic videos. We achieve an F1-score of 95.67% on the educational and adult videos set and an F1-score of 94% on our test subset for the pornographic class.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n A Clustering-Based Method for Automatic Educational Video Recommendation Using Deep Face-Features of Lecturers.\n \n \n \n \n\n\n \n Mendes, P. R. C.; Vieira, E. S.; Guedes, Á. L. V.; Busson, A. J. G.; and Colcher, S.\n\n\n \n\n\n\n . 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 6 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n\n\n\n
\n
@article{mendes_clustering-based_2020,\n\ttitle = {A Clustering-Based Method for Automatic Educational Video Recommendation Using Deep Face-Features of Lecturers},\n\turl = {http://arxiv.org/abs/2010.04676},\n\tabstract = {Discovering and accessing specific content within educational video bases is a challenging task, mainly because of the abundance of video content and its diversity. Recommender systems are often used to enhance the ability to find and select content. But, recommendation mechanisms, especially those based on textual information, exhibit some limitations, such as being error-prone to manually created keywords or due to imprecise speech recognition. This paper presents a method for generating educational video recommendation using deep face-features of lecturers without identifying them. More precisely, we use an unsupervised face clustering mechanism to create relations among the videos based on the lecturer's presence. Then, for a selected educational video taken as a reference, we recommend the ones where the presence of the same lecturers is detected. Moreover, we rank these recommended videos based on the amount of time the referenced lecturers were present. For this task, we achieved a {mAP} value of 99.165\\%.},\n\tjournaltitle = {{arXiv}:2010.04676 [cs]},\n\tauthor = {Mendes, Paulo R. C. and Vieira, Eduardo S. and Guedes, Álan L. V. and Busson, Antonio J. G. and Colcher, Sérgio},\n\turlyear = {2021},\n\tyear = {2020},\n\teprinttype = {arxiv},\n\teprint = {2010.04676},\n\tkeywords = {Computer Science - Machine Learning, Computer Science - Multimedia},\n}\n\n
\n
\n\n\n
\n Discovering and accessing specific content within educational video bases is a challenging task, mainly because of the abundance of video content and its diversity. Recommender systems are often used to enhance the ability to find and select content. But, recommendation mechanisms, especially those based on textual information, exhibit some limitations, such as being error-prone to manually created keywords or due to imprecise speech recognition. This paper presents a method for generating educational video recommendation using deep face-features of lecturers without identifying them. More precisely, we use an unsupervised face clustering mechanism to create relations among the videos based on the lecturer's presence. Then, for a selected educational video taken as a reference, we recommend the ones where the presence of the same lecturers is detected. Moreover, we rank these recommended videos based on the amount of time the referenced lecturers were present. For this task, we achieved a mAP value of 99.165%.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Future Vision of Interactive and Intelligent TV Systems using Edge AI.\n \n \n \n \n\n\n \n Guedes, Á. L.; Busson, A. J.; Navarro, J. P.; and Colcher, S.\n\n\n \n\n\n\n . 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n \n doi\n \n \n\n \n link\n \n \n\n bibtex\n \n\n \n\n \n \n \n 6 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{guedes_future_2020,\n\ttitle = {Future Vision of Interactive and Intelligent {TV} Systems using Edge {AI}},\n\tissn = {2446-9246},\n\turl = {https://www.set.org.br/ijbe/ed6/Artigo4.pdf},\n\tdoi = {http://dx.doi.org/10.18580/setijbe.2020.4},\n\tauthor = {Guedes, Álan L.V. and Busson, Antonio J.G. and João Paulo Navarro and Colcher, Sérgio},\n\turlyear = {2021},\n\tyear = {2020},\n}\n\n
\n
\n\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n A Cluster-Matching-Based Method for Video Face Recognition.\n \n \n \n \n\n\n \n Mendes, P. R. C.; Busson, A. J. G.; Colcher, S.; Schwabe, D.; Guedes, Á. L. V.; and Laufer, C.\n\n\n \n\n\n\n In
Proceedings of the Brazilian Symposium on Multimedia and the Web, pages 97–104, 2020. Association for Computing Machinery\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n \n Year\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n \n \n abstract \n \n\n \n \n \n 6 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n \n \n \n \n \n \n\n\n\n
\n
@inproceedings{mendes_cluster-matching-based_2020,\n\tlocation = {New York, {NY}, {USA}},\n\ttitle = {A Cluster-Matching-Based Method for Video Face Recognition},\n\tisbn = {978-1-4503-8196-3},\n\turl = {https://doi.org/10.1145/3428658.3430967},\n\tabstract = {Face recognition systems are present in many modern solutions and thousands of applications in our daily lives. However, current solutions are not easily scalable, especially when it comes to the addition of new targeted people. We propose a cluster-matching-based approach for face recognition in video. In our approach, we use unsupervised learning to cluster the faces present in both the dataset and targeted videos selected for face recognition. Moreover, we design a cluster matching heuristic to associate clusters in both sets that is also capable of identifying when a face belongs to a non-registered person. Our method has achieved a recall of 99.435\\% and a precision of 99.131\\% in the task of video face recognition. Besides performing face recognition, it can also be used to determine the video segments where each person is present.},\n\tpages = {97--104},\n\tbooktitle = {Proceedings of the Brazilian Symposium on Multimedia and the Web},\n\tpublisher = {Association for Computing Machinery},\n\tauthor = {Mendes, Paulo Renato C. and Busson, Antonio José G. and Colcher, Sérgio and Schwabe, Daniel and Guedes, Álan Lívio V. and Laufer, Carlos},\n\turlyear = {2021},\n\tyear = {2020},\n\tkeywords = {Clustering, Deep learning, Face recognition},\n}\n\n
\n
\n\n\n
\n Face recognition systems are present in many modern solutions and thousands of applications in our daily lives. However, current solutions are not easily scalable, especially when it comes to the addition of new targeted people. We propose a cluster-matching-based approach for face recognition in video. In our approach, we use unsupervised learning to cluster the faces present in both the dataset and targeted videos selected for face recognition. Moreover, we design a cluster matching heuristic to associate clusters in both sets that is also capable of identifying when a face belongs to a non-registered person. Our method has achieved a recall of 99.435% and a precision of 99.131% in the task of video face recognition. Besides performing face recognition, it can also be used to determine the video segments where each person is present.\n
\n\n\n
\n\n\n
\n
\n\n \n \n \n \n \n \n Avanços do middleware Ginga para TV 2.5.\n \n \n \n \n\n\n \n Álan L. V. Guedes; and Sérgio Colcher\n\n\n \n\n\n\n . 2020.\n
\n\n
\n\n
\n\n
\n\n \n \n Paper\n \n \n\n \n\n \n link\n \n \n\n bibtex\n \n\n \n\n \n \n \n 5 downloads\n \n \n\n \n \n \n \n \n \n \n\n \n \n \n\n\n\n
\n
@article{alan_l_v_guedes_avancos_2020,\n\ttitle = {Avanços do middleware Ginga para {TV} 2.5},\n\tissn = {1980-2331},\n\turl = {https://set.org.br/news-revista-da-set/artigo-news-revista-da-set/avancos-do-middleware-ginga-para-tv-2-5/},\n\tauthor = {{Álan L. V. Guedes} and {Sérgio Colcher}},\n\tyear = {2020},\n}
\n
\n\n\n\n
\n\n\n\n\n\n