An undergraduate research mentoring model in digital signal and image processing

International Journal of Electrical Engineering Education, Apr 2004 by Abdel-Qader, Ikhlas

Abstract

This work presents a research-mentoring framework to train undergraduate students in a hands-on setting in the area of digital signal and image processing (DSIP). It is intended to encourage undergraduate students to pursue research and to add to their knowledge in information technology and enhance their active-learning and problem-solving skills.

Keywords digital signal/image processing; undergraduate research

Recent advances in information technology are allowing more applications for digital signal and image processing (DSIP) to be added to the list. These applications include entertainment industry (video on demand and video games), medical (tele-medicine, medical instruments and different modalities of medical imaging), security and surveillance systems (object/face recognition and activity recognition), military applications (target recognition and/or detection), Communications (wireless and multimedia communications), and different industries (quality control, robot vision, intelligent transportation systems and vehicles, and nondestructive testing). These examples and many others involve digital information processing, compression, storage, visualization, analysis, and transport.

With the continued advancements in hardware, software, and the Internet, and with the rapid developments of information technology, the need for people with experience and research abilities in DSIP will grow exponentially. Furthermore, the competition around the globe is at its highest for advancements in DSIP applications. For example, digital signal processors (DSP) are the most rapidly expanding sector of the semiconductor market at a 30% annual growth rate since 1990. Since many industries have been making increasing demands for engineers with training in DSIP, educators must make available the necessary skills and opportunities to their students by expanding DSIP training.1 This training may include offering advanced DSIP courses and/or offering research experiences to undergraduate students. These offerings should be designed with the aim of encouraging undergraduate students to pursue advanced degrees and research careers. It also should aim to teach undergraduate students research through their senior level classes.

Recognizing the importance and intellectual challenges of this field, academic classes were initially taught at the graduate level and have more recently become core undergraduate classes for electrical and computer engineering programmes.2 The concepts, theory, and basic algorithms of digital signal and image processing have been established and applied for over thirty years.1 Typically, two approaches have been implemented in teaching these concepts to undergraduate students. It is accomplished through either classroom lectures and homework assignments, or a set of hands-on experiments with heavy reliance on programming. The model presented in this paper is different from both approaches because it is structured on research projects that accomplish the students' learning of DSIP in real time while engaged in a meaningful research experience that demonstrates the different stages of the research process. Consequently, motivating students to pursue graduate degrees, research careers, or advanced industry work in the DSIP area.

Furthermore, the project-based learning experience is different from the senior design experience in many aspects. Senior design is a capstone experience that reinforces all the concepts covered in the curriculum using a comprehensive design project. No discovery is involved. On the other hand, DSIP research projects involve new discoveries and concepts not typically covered in courses/curriculum. It involves working on problems/issues that have not been solved typically and it involves reading journal and conference research papers. In a senior design course, teams are mandatory, while in this research model individuality is expected. This allows for individual initiative and creativity. It also prevents the typical situation where team members did not work on the project equally. Additionally, students are required to take the senior design, while they elect to take the model discussed in this paper. The students favour the idea that they are different from their classmates and that they can elect to take advantage of work that is challenging. They all strived to produce scholarly work that could be published. Indeed, this issue appeared to motivate the students.

The digital signal and image processing curriculum at Western Michigan University (WMU) has been developed with strong emphasis on the theoretical and simulation aspects of signal processing. At the undergraduate level, it comprises a series of courses that provide a cohesive plan of study beginning with linear systems and random processes in the junior year followed by digital signal processing and digital communications in the senior year.3 For students with further interest in DSIP, independent study or faculty-directed research opportunities are available through both a project-based course ECE 490 and senior design capstone projects ECE 481/ECE 482 sequence. The project-based senior level course as well as a National Science Foundation (NSF) Research Experiences for Undergraduate (REU) supplemental grant has been utilized to develop the framework for the research-mentoring model discussed in this paper. Teaching students research turned out to be a challenging task, particularly as undergraduate students do not typically choose to conduct research. In other words, undergraduate students are used to being provided with information rather than seeking it.