PLA Information Engineering University

Unsupervised Change Detection (UCD) in Very High Resolution (VHR) Remote Sensing (RS) images remains to be a difficult challenge due to the inherent spatio-temporal complexity within data. Inspired by recent advancements in Visual Foundation Models (VFMs) and Contrastive Learning (CL), this research aims to develop CL methodologies to translate implicit knowledge in VFM into change representations, thus eliminating the need for explicit supervision. To this end, we introduce a Semantic-to-Change (S2C) learning framework for UCD in VHR RS images. Differently from existing CL methodologies that typically focus on learning multi-temporal similarities, we introduce a novel triplet learning strategy that explicitly models temporal differences, which are crucial to the CD task. Furthermore, random spatial and spectral perturbations are introduced during training to enhance robustness to temporal noise. In addition, a grid sparsity regularization is defined to suppress insignificant changes, and an IoU-matching algorithm is developed to refine the CD results. Experiments on three benchmark CD datasets demonstrate that the proposed S2C learning framework achieves significant improvements in accuracy, surpassing current state-of-the-art by over 31%, 9% and 23%, respectively. It also demonstrates robustness and sample efficiency, suitable for training and adaptation of various VFMs or backbone neural networks.

In the context of polar codes, belief propagation list (BPL) decoding has demonstrated a substantial enhancement in parallel decoding performance, achieving high throughput.Nevertheless, a performance gap still exists between the advanced BPL decoding and successive cancellation list (SCL) decoding methods.Moreover, existing bit-flipping strategies are inefficient in accurately identifying erroneous bit positions, leading to elevated computational complexity and limiting their practical applicability.This study introduces an optimized BPL decoding algorithm with dynamic flipping (OBPL-DF) aimed at bridging this performance gap while reducing computational demands.Initially, an efficient decoding scheme is proposed to further decrease computational complexity in practical scenarios.Subsequently, to improve the precision of error position detection, a partial cyclic redundancy check (CRC) code is employed on erroneous codewords.Finally, a dynamic flipping metric is developed within the bit-flipping strategy, allowing the selection of flipped positions to be guided by this novel metric rather than being confined to a predetermined set.Simulation results demonstrate that the OBPL-DF algorithm surpasses the performance of existing BPL flip (BPLF) decoding techniques and approaches that of enhanced SCL decoding, all while achieving significantly lower latency.

Droplet microlasers, as promising tools for biophotonics and biomedical sciences, have witnessed rapid advances due to their flexible reconfigurability, high sensitivity to stimuli, and label-free biosensing ability. However, designing these biosensors with simultaneously critical properties of low lasing threshold, high spectral purity, and ultimate sensitivity remains challenging. Here, we propose a versatile strategy to build liquid photonic molecules (LPMs) that combine all these features in a single device. We find that through tailoring the spectral Vernier overlap in size-mismatched droplets, this device enables single-mode lasing with a low threshold of ~610 nJ mm-2. The LPM lasers are engineered for dynamic tunability using a molecular isomerization strategy, which induces spectral mode hopping and thus yields a nearly ten-fold enhancement in spectral sensitivity over single droplets. Moreover, by leveraging the self-referenced intensity response of the LPM lasing modes, we demonstrate a three-orders-of-magnitude enhancement in biomolecular sensing, with a detection limit of 30 aM and a dynamic range spanning nine orders of magnitude. Our work offers exciting prospects for bio-integrated liquid sensors in diverse applications.

Abstract High-precision geomagnetic data is crucial for resource exploration and navigation. This paper proposes a reconstruction method integrating spatially adaptive mean compensation with patch-based sparse coding (MC-SC), which features a resolution-self-adaptive joint dictionary training strategy and a spatially adaptive MC module to preserve low-frequency backgrounds and recover high-frequency details. Experiments on three aeromagnetic survey regions show that MC-SC outperforms existing methods iterative back projection convolutional neural network (IBP and CNN) in reconstruction accuracy and detail fidelity. At a 50 m grid scale, MC-SC achieves a peak signal-to-noise ratio of 55.57 dB in Herat, significantly higher than IBP (52.89 dB) and CNN (52.64 dB). Moreover, under noisy conditions (up to 10 nT Gaussian noise), MC-SC exhibits superior robustness with slower performance degradation and more concentrated error distributions. The method provides a reliable solution for high-resolution geomagnetic data reconstruction from sparse samples, with strong potential for navigation and mapping applications.

Understanding information on the regional water–energy–food system pressure (WEFSP) is crucial for ensuring resource security and promoting sustainable regional development. Existing studies often lack a focus on water quality issues, which cannot fully reveal the current situation of WEFSP. This study incorporated the grey water footprint as a measurement indicator to integrate water quality into the WEF nexus, re-examining the WEFSP across 30 Chinese provinces from 2006 to 2020. The spatiotemporal evolutionary characteristics of the WEFSP were characterized using Standard Deviation Ellipse (SDE) and Kernel Density Estimation (KDE). Furthermore, the GeoDetector method was employed to identify the key driving factors and their interactive effects. The results revealed that (1) China’s WEFSP initially increased and then decreased, and the WEFSP changes the most during the five-year plan transition period. The energy subsystem was under the greatest pressure, while water quality scarcity caused by pollution was the dominant driver of pressure within the water subsystem. (2) Spatially, the WEFSP exhibited an east-high and west-low pattern, with the center of gravity of the WEFSP mainly located in Anhui and Henan provinces, and during the study period, it experienced two stages of transfer: from northwest to southeast and vice versa. (3) The explanatory power of driving factors for the spatial heterogeneity of the WEFSP exhibited dynamic variability. The most influential factor shifted from annual average precipitation to per capita consumption expenditure. Significant interactive effects were identified among factors, all demonstrating either bilateral or nonlinear enhancement. These findings provide a comprehensive insight into the current state of WEFSP and the influence of external factors, offering a scientific basis for formulating targeted resource management strategies to ensure the security of the WEF nexus.

Purpose To develop a digitized integrated feature-based interpretable machine learning classification model to accurately recognize complex thyroid nodules while efficiently diagnosing conventional thyroid nodules (thyroid nodules with typical benign or malignant US features). Materials and Methods Thyroid US images depicting pathologically confirmed nodules were retrospectively collected from seven medical centers in China (January 2011-December 2021). An interpretable classification model consisting of two independent masks was developed and defined as "UltraMC." The front-end network was trained to identify conventional thyroid nodules using four digitized features, and the back-end network collected nodules classified as benign in the previous framework for secondary analysis to clarify their final diagnosis. UltraMC performance was evaluated using accuracy, sensitivity, specificity, and confusion matrices. Results The total dataset included 73 826 patients with thyroid US images (mean age, 45.56 years ± 11.21 [SD]; 54 398 female). Diagnostic accuracy of the front-end network for detecting conventional thyroid nodules was 92.9% (13 718 of 14 765), and accuracy of the back-end network for classifying mummified thyroid nodules (MTNs) was 88.5% (652 of 737). The overall diagnostic accuracy of the US MTN classification model (UltraMC) was 91.8% (14 228 of 15 502). The areas under the receiver operating characteristic curve of the front-end network and UltraMC in identifying conventional thyroid nodules were 0.98 (95% CI: 0.98, 0.98) and 0.96 (95% CI: 0.96, 0.97), respectively. Conclusion The proposed two-layer interpretable classification model achieved high diagnostic accuracy for both conventional and mummified thyroid nodules. These findings demonstrate that digitized US features integrated into a white box framework can effectively support classification of complex thyroid nodules. Keywords: Ultrasound, Head/Neck, Thyroid, Diagnosis, Convolutional Neural Network (CNN), K-Means, Random Forest, Thyroid Nodule, Interpretable, Digital, Mummified Thyroid Nodules Supplemental material is available for this article. © RSNA, 2026.