IEEE Signal Processing Magazine

Signal processing is an engineering discipline known to involve abstract and complex concepts. Curriculum development should be informed by an understanding of the most critical and challenging learning in the field. Threshold concept theory and threshold capability theory provide a framework describing the features of the most critical and challenging learning in any discipline.

The effectiveness of teaching digital signal processing (DSP) can be enhanced by reducing lecture time devoted to theory and increasing emphasis on applications, programming aspects, visualization, and intuitive understanding. An integrated approach to teaching requires instructors to simultaneously teach theory and its applications in storage and processing of audio, speech, and biomedical signals.

Despite the impressive technological strides made over the years, human lives still depend very much on the natural environment. Fortunately, technology can now be used to help address critical environmental concerns in air quality, soil condition, and weather events.

Ten years ago, the world marveled at the ability of social media technology to assist an entire region in its pursuit of democracy. As I write this column days after the U.S. Presidential Inauguration, the world this time is overwhelmingly appalled by the role that same technology played in a violent attempt to overturn democracy. Those who decried the shutdown of access to social media desperately implemented by authoritarian regimes applauded similar restrictions implemented by tech companies in a quest to forestall additional violence.

This past summer, Prof. Robert Heath Jr. IEEE Signal Processing Magazine’s (SPM’s) former editor-in-chief, stressed to me how important it is to include a strong team of scientists on the magazine’s editorial board. It is especially important for area editors and members of the senior editorial board, but also associate editors for columns and forums as well as the e-Newsletter.

Many problems in signal processing [e.g., filter bank design, independent component analysis (ICA), beamforming design, and neural network training] can be formulated as optimization over groups of transformations that depend continuously on real parameters (Lie groups). Such problems are usually tackled in two ways: using a constrained optimization procedure or using some parameterization to transform them into unconstrained problems.

The old adage "you are what you wear" is taking on an entirely new meaning as smart watches, fitness trackers, and a rapidly expanding array of other wearable devices flood onto the market, enabling users to monitor their exercise progress, retrieve critical health data, and accomplish a wide range of other useful and informative tasks.

Deep neural networks provide unprecedented performance gains in many real-world problems in signal and image processing. Despite these gains, the future development and practical deployment of deep networks are hindered by their black-box nature, i.e., a lack of interpretability and the need for very large training sets. 

A little over a century and a half ago, Victor Hugo wrote “Il n’y a ni mauvaises herbes ni mauvais hommes. Il n’y a que de mauvais cultivateurs,” which translates to “there are no weeds and no bad men. There are only bad cultivators.” These two sentences provide a stark reminder of the heavy responsibility we all bear, as parents, educators, mentors, members of professional societies, and citizens of states, nations, and earth. Indeed, arguably our main goal as a professional society is to help develop our human capital. Everything else flows from there.

Enabling autonomous driving (AD) can be considered one of the biggest challenges in today?s technology. AD is a complex task accomplished by several functionalities, with environment perception being one of its core functions. Environment perception is usually performed by combining the semantic information captured by several sensors, i.e., lidar or camera. The semantic information from the respective sensor can be extracted by using convolutional neural networks (CNNs) for dense prediction. In the past, CNNs constantly showed stateof-the-art performance on several vision-related tasks, such as semantic segmentation of traffic scenes using nothing but the red-green-blue (RGB) images provided by a camera.