Loading...
Publication
OpenCV Studio
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 22.66 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A visão computacional tem gerado cada vez mais interesse devido às suas aplicações no campo da
inteligência artificial. Para suportar estes sistemas que utilizam a visão computacional, surgiram
bibliotecas que visam facilitar a implementação de comportamentos e funcionalidades, como é o
caso da biblioteca OpenCV. Esta é uma biblioteca complexa, constituída por diversos módulos
que disponibilizam inúmeros métodos, funções e algoritmos.
Foi desenvolvida uma plataforma low-code, intitulada OpenCV Studio, que tem como prin cipal objetivo elevar o nível de abstração da biblioteca OpenCV a fim de permitir o rápido
desenvolvimento de sistemas no campo da visão computacional de uma forma simples e intuitiva.
A avaliação da plataforma desenvolvida consistiu na realização de quatro tarefas por três par ticipantes. Estas realizadas em programação tradicional, recorrendo à linguagem de programação
C++, e posteriormente recorrendo à plataforma desenvolvida. Na maioria das tarefas verificou-se
uma diminuição de cerca de 50% no tempo necessário à implementação e conclusão das tarefas
através da plataforma low-code.
Durante a avaliação foram recolhidos os comentários e as sugestões dos participantes, através
da metodologia Think Aloud. Os participantes responderam também a um formulário SUS, para
se obter uma classificação de usabilidade da plataforma.
Ambas as avaliações apresentaram resultados positivos. Os participantes consideraram a pla taforma intuitiva e fácil de utilizar. A avaliação quantitativa, utilizando a System Usability Scale,
confirmou estes resultados apresentando uma pontuação de 84 pontos.
Computer vision has been generating increasing interest due to its applicability in the field of artificial intelligence. To support systems that use computer vision, libraries have emerged aimed at facilitating the implementation of behaviors and functionalities, such as the OpenCV library. This is a complex library, consisting of several modules that provide numerous methods, functions, and algorithms. A low-code platform, called OpenCV Studio, has been developed with the main goal of raising the level of abstraction of the OpenCV library, enabling the rapid development of computer vision systems in a simple and intuitive way. The evaluation of the developed platform consisted of completing four tasks by three partici pants. These tasks were first carried out using traditional programming in C++, and then using the developed platform. In most cases, the time required to implement and complete the tasks using the low-code platform was reduced by approximately 50%. During the evaluation, participants’ comments and suggestions were collected using the Think Aloud methodology. Participants also completed a SUS questionnaire to assess the platform’s usability. Both evaluations yielded positive results. The participants found the platform intuitive and easy to use. The quantitative evaluation, using the System Usability Scale, confirmed these findings with a score of 84 points.
Computer vision has been generating increasing interest due to its applicability in the field of artificial intelligence. To support systems that use computer vision, libraries have emerged aimed at facilitating the implementation of behaviors and functionalities, such as the OpenCV library. This is a complex library, consisting of several modules that provide numerous methods, functions, and algorithms. A low-code platform, called OpenCV Studio, has been developed with the main goal of raising the level of abstraction of the OpenCV library, enabling the rapid development of computer vision systems in a simple and intuitive way. The evaluation of the developed platform consisted of completing four tasks by three partici pants. These tasks were first carried out using traditional programming in C++, and then using the developed platform. In most cases, the time required to implement and complete the tasks using the low-code platform was reduced by approximately 50%. During the evaluation, participants’ comments and suggestions were collected using the Think Aloud methodology. Participants also completed a SUS questionnaire to assess the platform’s usability. Both evaluations yielded positive results. The participants found the platform intuitive and easy to use. The quantitative evaluation, using the System Usability Scale, confirmed these findings with a score of 84 points.
Description
Keywords
Visão computacional OpenCV Low-code Desenvolvimento visual Citizen developers Engenharia Informática . Faculdade de Ciências Exatas e da Engenharia