Towards Practical Computer Vision in Teaching and Learning of Image Processing Theories

This research to practice full paper presents a methodology that leads students towards innovative solutions to complex computer vision real problems. The cornerstone of teaching-learning process of image processing theories is based on strong interdisciplinary concepts (e.g. calculus, logic, data structures, perception, graphs, etc). However, just this theoretical background is not sufficient to reach the right practical understanding of how each type of technique interferes in an image considering real problems. To address this issue, we proposed to focus on the application of the image processing theories in a real industrial problem, through a computer vision pipeline. To do so, we used the concept of non-zero-sum game where both players win (e.g. students and industry). To evaluate the effectiveness of our methodology, students answered a questionnaire about their satisfaction. Based on these answers we map how much it helps the undergraduate students to understand the implications of different image processing concepts in a real environment. We also analyzed the satisfaction rate and the students' scores to obtain a well defined analysis. Our methodology presents a strong level of satisfaction according to the students, reaching up to 91% of acceptance rate.