Francisco de Assis Zampirolli

Distance Transform Separable by Mathematical Morphology in GPU

The Distance Transform DT is one of the classical operators in image processing, and can be used in Pattern Recognition and Data Mining, and there is currently a great demand for efficient parallel implementations on graphics cards, known as GPU. This paper presents simple and effective ways to implement the DT using decompositions of erosions with structuring functions implemented on GPU. The DT is equivalent to a morphological erosion of the binary image by a specific structuring function. However, this erosion can be decomposed by a sequence of erosions using small structuring functions. Classical and efficient algorithms of the DT are implemented on CPU. New 1D and 2D algorithms are implemented on GPU, using decomposition of structuring functions, inspired by implementations of convolution filters. All the GPU implementations used in this paper are known as brute-force, and even then present excellent results, comparable to the best CPU algorithms, which might contribute to future applications in image processing.

A smarter fire sprinkler

The following describes an application model of a robotic fire guard, an automated fire extinguishing system for detection and localized handling of potential outbreaks using an Infra-Red (IR) camera and a turret mounted dispenser. The acquired IR image is processed and the sources of heat isolated, the turret is successively aimed towards the fire by a targeting control system before releasing the fire suppressant. A proof of concept has been built using retail components and a commercial viability study has been performed pointing out its market potential.

Evaluation process for an introductory programming course using blended learning in engineering education

The offer of the same course simultaneously for thousands of students, whether in classroom teaching or in distance learning, requires standardization to compare performance between classes and to evaluate the success of the teaching–learning processes. This can be made easier through digital technologies. This study aims to present an Evaluation Process (EP) designed for an Introduction to Programming course (IP) for students of the Bachelors Degree in Science and Technology (Engineering Education) at the Federal University of ABC. Initially, the course scenario is presented in the classroom modality, Face‐to‐Face (IP‐FF), offered to about 2,000 students every year, identifying the problems, and then, describing the EP designed and applied in Blended Learning classes (IP‐BL). This includes the use of software tools developed for this purpose. The two scenarios (IP‐FF and IP‐BL) are compared using data from all the classes of the previous 9 years.

A Fast CUDA-Based Implementation for the Euclidean Distance Transform

In Image Processing efficient algorithms are always pursued for applications that use the most advanced hardware architectures. Distance Transform is a classic operation for blurring effects, skeletonizing, segmentation and various other purposes. This article presents two implementations of the Euclidean Distance Transform using CUDA (Compute Unified Device Architecture) in GPU (Graphics Process Unit): of the Meijsters Sequential Algorithm and another is a very efficient algorithm of simple structure. Both using only shared memory. The results presented herein used images of various types and sizes to show a faster run time compared with the best-known implementations in CPU.

An automatic generator and corrector of multiple choice tests with random answer keys

In the 21st century student-teacher communication in high schools has been increasingly computer-mediated. This brings a strong demand for experienced professionals worldwide, preferably graduated in Information Technology. Until the 19th century the opportunity to attend university courses was a privilege of a few undergraduates, so that the student-teacher relationship was highly personalized. In the 20th century the access to universities increased intensively in many countries, which is a very positive fact. However, it seriously compromised the personal student-teacher communication. This article presents an innovative solution to simplify generation and correction of multiple choice questions (MCQ) in which reliance on the results is wanted. With our program we obtain different issues of the same test with non-coinciding answer keys. Therefore, we offer an original, trustworthy and practical way to evaluate learning in courses attended by a large number of students. Our program generates a LTEX file and its compiled PDF file containing the individual exam of each student. The exam consists of a front page followed by a list of MCQ and optionally dissertation questions. The front page can be used as an answer sheet of the MCQ, and it has a layout that includes a header. One can opt for individual exams of which the answer keys to the MCQ is unique to each student. All these options are set in a separate configuration file. Our program uses a folder structure to organize classes of students in CSV files and databank of questions in TXT files. The students have to fill out their answer sheets of the front page, which will already contain both the student’s name and her/his id number in case of individual tests. After the exam the test front pages of a class can all be scanned into a single PDF file that our software will read to perform the automatic correction. The final scores are stored in a file that contains each student’s ID followed by the paired letters given.answer/answer.key of each single question. This file is in CSV format, which is both universally readable and writable by most spreadsheet programs. Our source codes are written in Python programming language. Our program is freely available on the Internet, where users can download the code in Python together with the folder structure and sample files to generate the exams. It has been intensively used at the Federal University of ABC in Brazil, both in classrooms and for distance learning courses, and also in simulations with 6772 tests with excellent results comparable with commercial products. This paper presents a software generator of MCQ which not only issues different versions of the same test, but also performs a fast correction of them all, and exports the results to a CSV file. Thanks to the facilities offered by automatic correction of MCQ the student-teacher communication can concentrate on the essential demands and become feasible again.In the 21st century student-teacher communication in high schools has been increasingly computer-mediated. This brings a strong demand for experienced professionals worldwide, preferably graduated in Information Technology. Until the 19th century the opportunity to attend university courses was a privilege of a few undergraduates, so that the student-teacher relationship was highly personalized. In the 20th century the access to universities increased intensively in many countries, which is a very positive fact. However, it seriously compromised the personal student-teacher communication. This article presents an innovative solution to simplify generation and correction of multiple choice questions (MCQ) in which reliance on the results is wanted. With our program we obtain different issues of the same test with non-coinciding answer keys. Therefore, we offer an original, trustworthy and practical way to evaluate learning in courses attended by a large number of students. Our program generates a LATEX file and its compiled PDF file containing the individual exam of each student. The exam consists of a front page followed by a list of MCQ and optionally dissertation questions. The front page can be used as an answer sheet of the MCQ, and it has a layout that includes a header. One can opt for individual exams of which the answer keys to the MCQ is unique to each student. All these options are set in a separate configuration file. Our program uses a folder structure to organize classes of students in CSV files and databank of questions in TXT files. The students have to fill out their answer sheets of the front page, which will already contain both the students name and her/his id number in case of individual tests. After the exam the test front pages of a class can all be scanned into a single PDF file that our software will read to perform the automatic correction. The final scores are stored in a file that contains each students ID followed by the paired letters given.answer/answer.key of each single question. This file is in CSV format, which is both universally readable and writable by most spreadsheet programs. Our source codes are written in Python programming language. Our program is freely available on the Internet, where users can download the code in Python together with the folder structure and sample files to generate the exams. It has been intensively used at the Federal University of ABC in Brazil, both in classrooms and for distance learning courses, and also in simulations with 6772 tests with excellent results comparable with commercial products. This paper presents a software generator of MCQ which not only issues different versions of the same test, but also performs a fast correction of them all, and exports the results to a CSV file. Thanks to the facilities offered by automatic correction of MCQ the student-teacher communication can concentrate on the essential demands and become feasible again.

Report of a Distance Learning course of specialization in Information Technology at a Brazilian public university

Despite the growing demand by Brazilian companies for well-qualified professionals in Information Technology (IT), the Brazilian educational system has not been able to meet this demand in a satisfactory way, especially the quality of vocational training question. One way to address this training problem has been the creation of specialized postgraduate courses in IT area focused on students already graduated and working in the IT market. To facilitate access to the course for these working students we opted for offering a Distance Learning model through a Web-based platform. Hence, this paper describes the structure and the experiments conducted over three offers of this course, in 2010, 2012 and 2014, which sought to resolve and measure various problems associated with Distance Learning, such as the sizing and scalability of teachers and tutors in relation to the number of students. Another problem tackled was student isolation, which was dealt with proactive attitudes and motivating actions by the tutors and teachers with tools such as chat, forums, FAQ and e-mails. The problem of “student authentication” was overcome with inperson periodical evaluations, plus continuous monitoring of the monograph thesis development by her/his supervising teacher and a final oral defense of the thesis.

An Optimal Cleaning Robot for all Kinds of Reflection Pools

In this paper we introduce the first part of a long-term project devoted to building an autonomous robot that is optimal at cleaning any kind of reflection pool. In our context, optimization refers to both minimal energy consumption and shortest time spent on cleaning the whole pool effectively. This means, the robot is programmed not to leave a single spot need either brushing or dirt suction. Our robot differs from other automata especially because it resorts to a unique combination of mathematical, computational and automative strategies that involve Plane Geometry, Complex Analysis and Genetic Algorithms.