Final Programme
September 30, 2022.
09.00-13.00 hrs. [Chile Continental Time. GMT -04.00]
| Organized by: | Núcleo de Inteligencia Artificial y Data Science Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte | |
| Sponsored by: | The Chilean Association for Pattern Recognition. | |
Chairs:
| Claudio Meneses Villegas (cmeneses@ucn.cl) Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte. | Diego Urrutia Astorga (durrutia@ucn.cl) Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte. |
| Date and Time: | Friday, September 30, 2022. 09 - 13 hrs. [Chile Continental Time. GMT -04.00] | |
Short description:
| The Workshop on Data and Knowledge Engineering (WDKE) is a space for the dissemination of scientific and professional academic activity in the area of data and knowledge enginnering, including invited talks and short presentations of research in progress works. | |
| Keynote speakers: |
PhD in Computer Science, University of Warwick. Senior Lecturer in Computer Science in the Department of Computer Science at City, University of London
Centro de Investigación de Estudios Avanzados del Maule, Laboratorio de Investigaciones Tecnológicas en Reconocimiento de Patrones (www.litrp.cl), Universidad Católica del Maule | |
Final Programme:
| 09.00-10.00 hrs. Invited Talk #1 Image segmentation of HeLa cells with HI (Human Intelligence) and AI (Artificial Intelligence), Speaker: Dr. Carlos Reyes-Aldasoro 10.00-11.00 hrs. Invited Talk #2 From synthetic palm vein imaging to large-scale biometric recognition, Speaker: Dr. Ruber Hernández García 11.00-13.00 hrs. Session of Short Presentations
| |
Invited Talk #1
Friday 30/09/2022 09.00-10.00 am.
Title: Image segmentation of HeLa cells with HI (Human Intelligence) and AI (Artificial Intelligence)
Speaker: Dr. Constantino Carlos Reyes Aldasoro Affiliation: PhD in Computer Science, University of Warwick. Senior Lecturer in Computer Science in the Department of Computer Science at City, University of London
Short Abstract. Image segmentation of biomedical images has a long history and numerous algorithms for numerous applications have been developed. In recent years, deep learning techniques, like convolutional neural networks, have provided very good results, in some cases better than those of “traditional” algorithms and in other cases even better than human experts. In this presentation, we will explore the journey (which may seem a Magical Mystery Tour) of applying HI (Human Intelligence) and AI (Artificial Intelligence) to the segment HeLa cells observed with Electron Microscopy. A comparison and a discussion between techniques will be presented together with experiences along the journey. | |
Email: Constantino-Carlos.Reyes-Aldasoro@city.ac.uk
| Short Bio. Dr. Reyes-Aldasoro is an interdisciplinary scientist with interest in Computer Science, Engineering and Life Sciences, in particular the computational analysis of data from Cancer, Microcirculation and Inflammation observed with a variety of acquisition techniques, from Electron Microscopy to Magnetic Resonance Imaging. He has specialised in Biomedical Image Analysis and has more than 20 years of experience in the area. Dr Reyes-Aldasoro is author of a book (Biomedical Image Analysis Recipes in MATLAB: For Life Scientists and Engineers, 2015, Wiley), numerous peer-reviewed journal and conference papers, edited several conference proceedings and special issues in journals like Medical Image Analysis. He is academic editor of PLoS ONE, Immuno-Informatics, Journal of Imaging and previously of Oncology News. He has served in the Executive Committees of the British Association for Cancer Research, IET Vision and Imaging Technical Network, Medical Image Understanding and Analysis Conference.
He has a degree in Mechanical and Electrical Engineering (UNAM, Mexico), MSc in Electrical Engineering (Imperial College) and PhD in Computer Science (Warwick) and worked as a Research Associate and Fellow at the Department of Surgical Oncology, School of Medicine of The University of Sheffield. He is currently Senior Lecturer in Computer Science in the Department of Computer Science at City, University of London where he is currently supervising 6 doctoral students and has supervised 6 to completion. |
Invited Talk #2
Friday 30/09/2022 10.00-11.00 am
Title: From synthetic palm vein imaging to large-scale biometric recognition
Speaker: Dr. Ruber Hernández García Affiliation: Centro de Investigación de Estudios Avanzados del Maule, Laboratorio de Investigaciones Tecnológicas en Reconocimiento de Patrones (www.litrp.cl), Universidad Católica del Maule
Short Abstract. Palm vein recognition is an emerging biometric technique with several advantages, especially in terms of security against forgery, which has gained the attention of the research community. However, collecting large-scale biometric datasets is a challenging task because of restrictions on time, security, and cost. Publicly available databases have a reduced number of individuals and samples lacking detailed annotations on soft traits (e.g., gender, age, weight). Therefore, evaluating the scalability of developed methods on massive datasets is not feasible, and the influences of different attributes have been poorly investigated. This talk will dive deeply into the suitability of synthetic vein images generated to compensate for the urgent lack of publicly available large-scale datasets. Firstly, it will present an overview of recent research progress on palm vein recognition, from the basic background knowledge to vein anatomical structure, data acquisition, public database, and recognition approaches. Later, it will also examine state-of-the-art methods that have allowed the generation of vascular structures for biometric purposes. Finally, it will introduce a general flowchart for creating a synthetic palm vein database and a conceptual mathematical model to generate synthetic palm vein images, analyzing the performance of synthetic palm vein datasets for biometric recognition. | |
Email: rhernandez@ucm.cl
| Short Bio. Dr. Ruber Hernández-García (ORCID: 0000-0002-9311-1193, http://www.ruberhg.com) received the title in Engineering in Informatics Sciences from the University of Informatics Sciences, La Habana, Cuba, in 2007; M.S. degrees in Audiovisual Information Systems and in Applied Informatics from University of Malaga (Spain) and University of Informatics Sciences, in 2010 and 2011, respectively; and the Ph.D. degree in Computer Science from University of Informatics Sciences in 2014. Currently, he is Academic Researcher in Laboratory of Technological Research in Pattern Recognition (LITRP) and the Research Center for Advanced Studies of Maule (CIEAM), at the Universidad Católica del Maule, Chile. He currently head the research project FONDECYT Iniciación en Investigación 2022 No. 11220693 “End-to-end multi-task learning framework for individuals identification through palm vein patterns”, funded by the Chilean Government. He served as the reviewer for international journals and conferences, including IEEE Transactions on Pattern Analysis and Machine Intelligence, Expert Systems and Applications, IEEE Access, Journal of Supercomputing, Complex & Intelligent Systems, Sensors, Micromachines, and ICPRS. His research interests include video and image processing, pattern recognition, and biometric systems under parallel platforms. |
Short presentation #1
Friday 30/09/2022 11.00-11.20 hrs.
Title: Asociación Chilena de Reconocimiento de Patrones
Speaker: Dr. César A. Astudillo Académico, Universidad de Talca, Presidente ACHIRP
Short Abstract.
La Asociación Chilena de Reconocimiento de Patrones (ACHIRP) es una sociedad científica sin fines de lucro que está asociada a la International Association for Pattern Recognition (IAPR). En esta presentación se expondrá brevemente los objetivos de la ACHIRP, una descripción de los eventos que organizamos, los beneficios de pertenecer a nuestra comunidad y como unirse a nuestra sociedad. Académicos, estudiantes y profesionales del sector privado están cordialmente invitados.
| |
Email: castudillo@utalca.cl | Short Bio.
César Castillo es Ingeniero Civil Informático (Universidad de Concepción) y posee el grado de Doctor en Ciencia Computacional (Carleton University). Sus conocimientos técnicos, abarcan la ciencia computacional, inteligencia artificial y aprendizaje profundo. La ciencia que realiza incluye la creación de nuevos algoritmos de inteligencia artificial que permiten a las máquinas aprender de los datos. En otra dimensión científica, César integra equipos multidisciplinarios para resolver problemas de ingeniería que requieran inteligencia artificial. Algunos ejemplos de campos de aplicación en los que ha aportado soluciones incluyen Agronomía, Bio-informática, Economía, Electrónica, Química, Logística, Medio Ambiente, y Educación, entre otros. Por casi veinte años, César se ha desempeño como académico en la Universidad de Talca, y entre otros roles ha sido Director de Departamento de Ciencias de la Computación, y ha formado parte del Comité Académico del Doctorado en Sistemas de Ingeniería. Actualmente es el presidente de la Asociación Chilena de Reconocimiento de Patrones, entidad científica sin fines de lucro que busca la diseminación de dicho sub-campo de la inteligencia artificial. |
Short presentation #2
Friday 30/09/2022 11.20-11.40 hrs.
Title: An Empirical Evaluation of Supervised Learning Methods for Network Malware Identification Based on Feature Selection
Speaker: Carlos Manzano Munizaga Escuela de Ingeniería, Universidad Católica del Norte, Coquimbo, Chile.
Short Abstract.
Malware is a sophisticated, malicious, and sometimes unidentifiable application on the network. The classifying network traffic method using machine learning shows to perform well in detecting malware. In the literature, it is reported that this good performance can depend on a reduced set of network features. This study presents the evaluation of two statistical methods of reduction and selection of features in an Android network traffic dataset using six well-known supervised machine learning algorithms.
| |
Email: cmanzano@ucn.cl | Short Bio.
Carlos Manzano earned a master in information technology degree from UFSM, a master (c) computer engineering from UCN, and currently he is a PhD student at UFSM, Valparaíso, Chile. Also, he is a teacher in the School of Engineering at Universidad Católica del Norte, Coquimbo, Chile. His main research area is related to computer networks and distributed systems. |
Short presentation #3
Friday 30/09/2022 11.40-12.00 hrs.
Title: Modelos Predictivos en Plantas de Osmosis Reversa
Speaker: Mg. Diego Urrutia Astorga Profesor Asistente, Depto. de Ingeniería de Sistemas y Computación, UCN, Antofagasta, Chile.
Short Abstract.
En la presentación se mostrará el trabajo realizado en la aplicación de las primeras 5 fases del modelo de proceso CRISP-DM, sobre la problemática del control y monitoreo de datos operacionales en plantas de osmosis reversa. Se mostrarán detalles del proceso desarrollado, junto con los problemas y soluciones encontradas.
| |
Email: durrutia@ucn.cl | Short Bio.
Diego Urrutia Astorga es profesor tiempo completo del Departamento de Ingeniería de Sistemas y Computación de la UCN, sede Antofagasta, Chile. El obtuvo el grado de Magíster en Ingeniería Informática y el título de Ingeniero de Ejecución en Computación e Informática, ambos desde la Universidad Católica del Norte. Sus intereses de investigación están relacionados con la Ingeniería de Software, la Computación de Alto Rendimiento y Big Data. |
Short presentation #4
Friday 30/09/2022 12.00-12.20 hrs.
Title: Explorando la detección de los alimentos mediante algoritmos de aprendizaje profundo.
Speaker: Ing. Juan Quispe Ticona Universidad Católica del Norte, Antofagasta, Chile.
Short Abstract.
Durante los últimos años las personas han demostrado tener una mayor preocupación sobre su dieta alimentaria, ya sea para prevenir enfermedades, para tratamientos médicos u otros. En comidas servidas en restaurantes, colegios o comedores públicos no es fácil identificar los ingredientes y/o información nutricional contenida en estas. Actualmente existen soluciones tecnológicas basadas en modelos de aprendizaje profundo que facilitan el registro y seguimiento de los alimentos consumidos a partir de una imagen. Teniendo en cuenta de que en ocasiones es posible que existan múltiples ítems de comidas servidas en un mismo plato, el análisis de los alimentos debe ser tratado como un problema de detección de objetos y no de clasificación. EfficientDet y YOLOv5 son algoritmos de detección de objetos que han demostrado una alta precisión y funcionamiento en tiempo real sobre datos de dominio general. Sin embargo, estos modelos no han sido evaluados y comparados sobre bases de datos públicas de alimentos. A diferencia de los objetos de dominio general, los alimentos poseen características más desafiantes propias de su naturaleza que elevan la complejidad de la detección. En este artículo, se realiza una comparativa del desempeño de EfficientDet y YOLOv5 sobre UNIMIB2016 y UECFood256. De los resultados obtenidos, podemos observar que YOLOv5 provee una diferencia significativa tanto en términos de precisión como de tiempo de respuesta en comparación con EfficientDet. Además, YOLOv5 obtiene un rendimiento superior al evidenciado en el estado-del-arte sobre UECFood256, logrando una mejora de más de un 4,95% en términos de mAP@0.5.
| |
Email: jqt002@alumnos.ucn.cl | Short Bio.
Juan Quispe Ticona es Licenciado en Ciencias de la Ingeniería y egresado de Ingeniería Civil en Computación e Informática de la Universidad Católica del Norte, Antofagasta, Chile. Su principal interés en la investigación es el desarrollo y la aplicación de algoritmos de inteligencia artificial en el dominio de la salud. |
Short presentation #5
Friday 30/09/2022 12.20-12.40 hrs.
Title: Academic Performance Evaluation Using Data Mining in Times of Pandemic
Speaker: Dr. Edgar Aurelio Taya Acosta Universidad Nacional Jorge Basadre Grohmann, Tacna, Perú.
Short Abstract.
This work focuses on studying the relationship that existed between the use of the learning management system (LMS) and the academic performance of the students of the Jorge Basadre Grohmann National University of Tacna-Perú. For this, we use the data provided by the LMS (access virtual classroom) and the university's academic management system (grades). For that, we perform various classification machine learning algorithms to predict academic performance with two classes SATISFACTORY or POOR where Gradient Boosted Trees algorithm had the best accuracy 91.79%. However, with three classes, SATISFACTORY, REGULAR AND POOR, Random Forest algorithm had the best accuracy of 89.26%
| |
Email: etayaa@unjbg.edu.pe | Short Bio.
Profesor Universitario de Pre-Grado y Posgrado de la Universidad Nacional Jorge Basadre Grohmann-Tacna de la Escuela Profesional de Ingeniería en Informática y Sistemas desde Julio 2001 hasta la actualidad (Dictado de cursos de Base de Datos, Ingeniería Web y Big Data con Analítica de Datos) Conferencista Nacional e Internacional - Profesor invitado en la: Universidad de Santiago de Compostela (España)-2003, Universidad de Boyacá (Colombia) – 2009, Universidad de Tarapacá (Chile) - 2017, 2019; Universidad Arturo Prat (Chile – 2018, Universidad Católica del Norte de Chile – 2019, Pontificia Universidad Católica de Chile – 2020, Universidad de Atacama (Chile) – 2020. Árbitro Internacional revisor de la Revista Ingeniare (indexada en Scielo-Chile), Árbitro Internacional revisor de la Revista Conciencia Tecnológica (indexada en Redalyc-México). |
Short presentation #6
Friday 30/09/2022 12.40-13.00 hrs.
Title: Social distancing detection by drone exploration
Speaker: Bach. Ruth de Jesus Ramírez Rejas Universidad Nacional Jorge Basadre Grohmann, Tacna, Perú.
Short Abstract.
Facing the criticality of the COVID 19 pandemic, we propose an artificial intelligence system with a modern approach detecting people and their social distancing in crowded places using thermal images obtained from the DJI Mavic 2 Enterprise Dual drone. We implement an algorithm that analyzes two types of images: color and thermal, to measure the distance between people. We used the Fast R-CNN neural network, the images with videos were extracted from the DJI Pilot application. The objective is to identify the distance between people. The results obtained show that the proposed algorithm is suitable for monitoring the city.
| |
Email: rramirezr@unjbg.edu.pe | Short Bio.
Bachiller en ciencias con mención en Informática y Sistemas Egresada de la Escuela Profesional en Ingeniería en Informática y Sistemas de la Universidad Nacional Jorge Basadre Grohmann del año 2020.
|
