This is our sixth and final issue of 2024. It is hard to believe that a year has gone by since our term as the new editorial team started in January. In our first year, in addition to our usual array of technical overviews and Society news, we addressed a number of topics of significance for our community in the hopes of starting a discussion.

I am writing this short note as I am about to board a plane to Abu Dhabi to join those of you who are attending the 2024 edition of the International Conference on Image Processing (ICIP 2024). The team organizing ICIP 2024 has put together an outstanding technical program that includes world-class plenary speakers discussing research and industrial trends.

Multichannel acoustic signal processing is a well-established and powerful tool to exploit the spatial diversity between a target signal and nontarget or noise sources for signal enhancement. However, the textbook solutions for optimal data-dependent spatial filtering rest on the knowledge of second-order statistical moments of the signals, which have traditionally been difficult to acquire.

“All models are wrong, but some are useful” - understanding “models” as analytical mathematical models, this aphorism, originating from George Box in 1976, motivates the synthesis of model-based and data-driven audio signal processing as the leitmotif of this special issue.

Multichannel acoustic signal processing is a well-established and powerful tool to exploit the spatial diversity between a target signal and nontarget or noise sources for signal enhancement. However, the textbook solutions for optimal data-dependent spatial filtering rest on the knowledge of second-order statistical moments of the signals, which have traditionally been difficult to acquire.

Communication base stations can achieve high-precision tracking and accurate classification for multiple extended targets in the context of integrated communication and sensing by transmitting wideband signal. However, the time resources of the base stations are often limited. In the time-division operation mode, part of the time resources must be reserved to guarantee communication performance, while the rest of the resources must be properly allocated for better multi-target sensing performance.

We consider a least absolute deviation (LAD) approach to the robust phase retrieval problem that aims to recover a signal from its absolute measurements corrupted with sparse noise. To solve the resulting non-convex optimization problem, we propose a robust alternating minimization (Robust-AM) derived as an unconstrained Gauss-Newton method.

This paper explores constrained non-convex personalized federated learning (PFL), in which a group of workers train local models and a global model, under the coordination of a server. To address the challenges of efficient information exchange and robustness against the so-called Byzantine workers, we propose a projected stochastic gradient descent algorithm for PFL that simultaneously ensures Byzantine-robustness and communication efficiency. 

Steganography is the art of covert communication that pursues the secrecy of concealment. In adaptive steganography, the most commonly used framework of steganography, the sender embeds a “secret message” signal within another “cover” signal with respect to a certain adaptive distortion function that measures the distortion incurred, contributing to the composite “stego” signal that resembles the cover, and the receiver extracts the “secret message” signal from the stego.