Title: Token-Wise Attention-Guided Semantic Quality Assessment for Compressed Visual Features

Authors: Shien Ke, Changsheng Gao, Hadi Amirpour, Zhihua Wang,  Xaoyan Sun

Event: QoMEX 2026, Cardiff, UK, June 29th – July 3rd, 2026

Abstract: In collaborative and distributed intelligent systems, compressed intermediate features are routinely transmitted and reused, making semantic quality assessment (SQA) crucial for reliable deployment. Recent compressed feature quality assessment (CFQA) benchmarks, however, show that conventional similarity measures often correlate poorly with downstream semantic utility and lack robustness across diverse feature codecs. In this paper, we propose a token-wise, attention-guided method for assessing the semantic quality of compressed features. First, motivated by the observation that many downstream heads normalize and process tokens largely independently, we assess quality at the token level. This token-wise formulation exploits the intrinsic correspondence between the original and reconstructed tokens while reducing cross-token interference. Second, since tokens contribute unequally to downstream task performance, we adopt an attention-guided aggregation scheme: we derive task-adaptive importance weights from DINOv2 self-attention and use them to pool token-wise quality predictions into a global semantic quality score. Third, to accommodate heterogeneous supervision across tasks, we cast CFQA as a regression problem and rescale classification-based rank targets to mitigate label imbalance. Experiments on the CFQA benchmark demonstrate that our method consistently improves PLCC and SROCC across three tasks and four codecs, yielding a practical, codec-agnostic quality interface for next-generation intelligent systems.

Title: EPS: Efficient Patch Sampling for Video Overfitting in Deep Super-Resolution Model Training

Authors: Yiying Wei, Hadi Amirpour, Jong Hwan Ko, and Christian Timmerer

Abstract: Leveraging the overfitting property of deep neural networks (DNNs) is trending in video delivery systems to enhance video quality within bandwidth limits. Existing approaches transmit overfitted super-resolution (SR) model streams for low-resolution (LR) bitstreams, which are used to reconstruct high-resolution (HR) videos at the decoder. Although these approaches show promising results, the huge computational costs of training a large number of video frames limit their practical applications. To overcome this challenge, we propose an efficient patch sampling method named EPS for video SR network overfitting, which identifies the most valuable training patches from video frames.

To this end, we first present two low-complexity Discrete Cosine Transform (DCT)-based spatial-temporal features to measure the complexity score of each patch directly. By analyzing the histogram distribution of these features, we then categorize all possible patches into different clusters and select training patches from the cluster with the highest spatial-temporal information. The number of sampled patches is adaptive based on the video content, addressing the trade-off between training complexity and efficiency.

Our method reduces the number of training patches by 75.00\% to 91.69\%, depending on the resolution and number of clusters, while preserving high video quality and greatly improving training efficiency. Our method speeds up patch sampling by up to 82.1$\times$ compared to the state-of-the-art patch sampling technique (EMT).

Title: Performance Evaluation of Privacy Models for Data Streams on the Edge

Authors: Ilir Murturi, Boris Sedlak, Reza Farahani, Schahram Dustdar

Venue: Internet Technology Letter

Abstract: Recent advances in edge computing enable data stream privacy enforcement directly on resource‐constrained devices, reducing latency and the exposure of sensitive information. In this paper, we extend and validate our previously proposed privacy‐enforcing framework, which allows high‐level privacy policies to be expressed as chains of triggers and transformations, executed at the edge. To assess its practical viability, we conduct a comprehensive performance profiling of multiple privacy models across heterogeneous edge hardware platforms. Six privacy‐model chains, ranging from basic face detection to combined face‐and‐person anonymization, are evaluated across three representative edge devices. Key performance metrics (i.e., execution time, CPU utilization, memory usage, and power consumption) are measured to inform optimal placement of privacy transformations. Our evaluation offers critical insights into the effectiveness of the privacy‐enforcing framework on resource‐constrained devices, thereby guiding practitioners in selecting suitable deployment targets for privacy‐preserving data stream analytics on the edge.

Paper title: AI-Assisted Energy-Efficient Multimedia Systems

Authors: Zoha Azimi

Venue:  (Doctoral Symposium) MMSys’26, The 17th ACM Multimedia System Conference, Hong Kong SAR, 4th – 8th April 2026

Abstract: Video streaming constitutes the majority of today’s Internet traffic and is expected to continue growing in scale, complexity, and environmental impact. As video systems, from encoding to delivery, consume substantial computational resources, their energy footprint has emerged as a critical challenge for both industry and research communities. At the same time, recent advances in Artificial Intelligence (AI) and Generative AI have improved video quality and adaptive streaming performance, yet often at the cost of increased computational load and higher energy consumption. This increase creates a growing need for streaming systems that not only deliver high Quality of Experience (QoE) but also minimize energy usage across heterogeneous devices and network infrastructures. The scope of this doctoral study is within the end-to-end video ecosystem, focusing on balancing the trade-off between energy consumption and user experience. It aims to investigate and develop intelligent frameworks that treat sustainability as a primary metric alongside performance, improving energy efficiency across the video lifecycle without compromising the viewer’s experience. We present three fundamental research questions to target the related challenges in the domain of sustainable multimedia systems.