In the context of lossy compression, Blau & Michaeli [1] adopt a mathematical notion of perceptual quality and define the information rate-distortion-perception function, generalizing the classical rate-distortion tradeoff. We consider the notion of universal representations in which one may fix a rate and an encoder then vary the decoder to achieve any point within a collection of distortion and perception constraints. We prove that the corresponding information-theoretic universal rate-distortion-perception function is operationally achievable in an approximate sense. Under MSE distortion, we show that the entire distortion-perception tradeoff of a Gaussian source can be achieved by a single encoder of the same rate asymptotically. We then characterize the achievable distortion-perception region for a fixed representation in the case of arbitrary distributions, and identify conditions under which the aforementioned results continue to hold approximately. Finally, we extend our notion of universality to the case where the rate is no longer fixed and additional bits can be sent at a second stage, generalizing the classical theory of successive refinement [2] with perception constraints. This motivates the study of practical constructions that are approximately universal across the RDP tradeoff, thereby alleviating the need to design a new encoder for each objective. We provide experimental results on MNIST and SVHN suggesting that on image compression tasks, the operational tradeoffs achieved by machine learning models with a fixed encoder suffer only a small penalty when compared to their variable encoder counterparts.

This paper studies the rate-distortion-perception (RDP) tradeoff for a Gaussian vector source coding problem where the goal is to compress the multi-component source subject to distortion and perception constraints. Specifically, the RDP setting with either the Kullback-Leibler (KL) divergence or Wasserstein-2 metric as the perception loss function is examined, and it is shown that for Gaussian vector sources, jointly Gaussian reconstructions are optimal. We further demonstrate that the optimal tradeoff can be expressed as an optimization problem, which can be explicitly solved. An interesting property of the optimal solution is as follows. Without the perception constraint, the traditional reverse water-filling solution for characterizing the rate-distortion (RD) tradeoff of a Gaussian vector source states that the optimal rate allocated to each component depends on a constant, called the water level. If the variance of a specific component is below the water level, it is assigned a zero compression rate. However, with active distortion and perception constraints, we show that the optimal rates allocated to the different components are always positive. Moreover, the water levels that determine the optimal rate allocation for different components are unequal. We further treat the special case of perceptually perfect reconstruction and study its RDP function in the high-distortion and low-distortion regimes to obtain insight to the structure of the optimal solution.

The waviness defect, a common manufacturing flaw in composite structures, can significantly impact the mechanical performance. This study investigates the effects of wrinkles on the ultimate load and failure modes of two Carbon Fiber Reinforced Composite (CFRC) laminates through compressive experiments and simulation analyses. The laminates have stacking sequences of [0]_10S and [45/0/−45/90/45/0/−45/0/45/0]_S. Each laminate includes four different waviness ratios (the ratio of wrinkle amplitude to laminate thickness) of 0%, 10%, 20% and 30%. In the simulation, a novel multiaxial progressive damage model is implemented via the user material (UMAT) subroutine to predict the compressive failure behavior of wrinkled composite laminates. This multiscale analysis framework innovatively features a 7 × 7 generalized method of cells coupled with stress-based multiaxial Hashin failure criteria to accurately analyze the impact of wrinkle defects on structural performance and efficiently transfer macro-microscopic damage variables. When any microscopic subcell within the representative unit cell (RUC) satisfies a failure criterion, its stiffness matrix is reduced to a nominal value, and the corresponding failure modes are tracked through state variables. When more than 50% fiber subcells fail in the fiber direction or more than 50% matrix subcells fail in the transverse or thickness direction, it indicates that the RUC has experienced the corresponding failure modes, which are the tensile or compressive failure of fibers, matrix, or delamination in the three axial directions. This multiscale model accurately predicted the load–displacement curves and failure modes of wrinkled composites under compressive load, showing good agreement with experimental results. The analysis results indicate that wrinkle defects can reduce the ultimate load-carrying capacity and promote local buckling deformation at the wrinkled region, leading to changes in damage distribution and failure modes.

With the rapid growth of the low-altitude economy and urban air mobility, ensuring the operational safety of low-altitude aerial vehicles in dense urban environments has become a critical engineering challenge, particularly under GPS loss or abnormal flight conditions that require emergency landing. Reliable real-time perception of feasible landing zones is therefore essential to support autonomous emergency landing decision-making for unmanned aerial vehicles and eVTOL aircraft. To address this challenge, this paper proposes LZSeg-PID, a real-time semantic segmentation framework with dedicated architectural and learning design for safety-critical aerial perception. The proposed method introduces a tri-branch architecture that explicitly decouples and jointly models detailed spatial information, global contextual semantics, and boundary-aware features. In addition, a boundary-aware multi-task learning strategy is developed, incorporating explicit boundary supervision and OHEM, to enhance small-object recognition and boundary discrimination under real-time constraints. The proposed framework is evaluated on three representative aerial benchmarks—UDD6, VDD, and UAVid—covering complex urban scenes with heterogeneous surface categories and scale variations. Experimental results demonstrate that LZSeg-PID achieves a favorable trade-o! between segmentation accuracy and inference e"ciency, consistently outperforming state-of-the-art lightweight segmentation baselines. The accuracy-oriented variant reaches 77.1% mIoU (62.6 FPS), 81.1% mIoU (43.2 FPS), and 69.3% mIoU (39.8 FPS) on the respective datasets, while the lightweight variant (7.6M parameters) attains 73.7% mIoU (136.7 FPS), 71.9% mIoU (101.1 FPS), and 67.6% mIoU (93.2 FPS). These results demonstrate the e!ectiveness and practical feasibility of the proposed framework for real-time onboard perception in low-altitude emergency landing scenarios.

Autonomous UAV landing in urban scenes requires a perception module that can identify candidate landing surfaces, reject structural and dynamic hazards, and express uncertainty before a downstream controller commits to a landing maneuver. This study reformulates UAV landing perception as a unified three-class landing-safety segmentation problem by relabeling UAVid, UDD6, and VDD into candidate landing area, structural obstacle, and critical hazard classes. The Landing3 dataset is constructed with an implementation-consistent relabeling protocol in which critical hazards override other labels and only sufficiently large connected components of source-specific candidate classes are retained as landing candidates. Two real-time segmentation models are evaluated under this unified task: PIDNet, a CNN-based multi-branch model with explicit boundary modeling, and SCTNet, a Transformer-guided model with semantic alignment for long-range context modeling. A post hoc conformal prediction (CP) module then converts softmax outputs into pixel-wise prediction sets, and the final candidate landing region is extracted only from pixels whose prediction set is the singleton candidate-landing class. Experiments compare the two models in terms of best validation mIoU, class-wise IoU, resolution-dependent accuracy–speed trade-offs, power-constrained FPS and complexity, qualitative candidate-area visualization, and CP-derived safe-area quality. A fixed-checkpoint sensitivity analysis further shows that the comparative model ranking remains stable across the tested connected-component threshold settings. SCTNet provides a lighter model and higher throughput under most tested resolutions and power limits, whereas PIDNet preserves higher safe-area recall, safe IoU, and spatial coherence after CP filtering. These results show that reliable UAV landing perception requires joint consideration of cross-dataset task definition, real-time model efficiency, and uncertainty-aware candidate-area extraction, rather than semantic segmentation accuracy alone.