Search Results (12)

Screen content images (SCIs) exhibit complex structural heterogeneity, rendering traditional statistics-based natural scene image quality assessment (NR-IQA) metrics ineffective. Although deep learning models achieve high prediction accuracy, their prohibitive computational demands preclude deployment in latency-sensitive industrial scenarios. While existing handcrafted lightweight SCI-IQA metrics reduce computational overhead, most rely on unsegmented global feature pooling or holistic edge statistics (e.g., edge histograms or Fisher vector coding), thereby diluting locally critical text-edge distortions in vast homogeneous backgrounds. To address this limitation, we propose an ultra-lightweight, deep-learning-free NR-IQA framework centered on human visual masking. Unlike existing lightweight methods, our approach explicitly employs dual-scale Canny edge operators to partition SCIs into edge-sensitive and flat background regions. Guided by this visual prior, structural degradations and micro-compression textures are extracted region-wise using Sobel gradients and uniform local binary patterns (LBPs) and aggregated with global Commission Internationale de I’Eclairage L*a*b*(CIELAB) color statistics into a compact 60-dimensional descriptor. A grid-search-optimized Support Vector Regression (SVR) maps these features to subjective quality scores. Extensive cross-validation on the SIQAD and SCID datasets demonstrates that our metric outperforms existing handcrafted lightweight SCI metrics and traditional NSS models, while achieving accuracy competitive with representative full-reference metrics. Consuming only 79.3 ms per image on a standard CPU, it offers a practical accuracy–efficiency trade-off for resource-constrained periodic quality monitoring. Full article

Strict minimum message length (SMML) is an information-theoretic coding principle that represents a continuous statistical model by a finite set of assertions and a partition of the sample space. We show that the SMML objective decomposes into assertion entropy and conditional cross-entropy, balancing the cost of identifying an assertion against the cost of encoding data under the assigned model. For any fixed partition, the optimal codepoint for each cell is the model distribution that minimises Kullback–Leibler (KL) divergence from the data distribution restricted to that cell. Using the local Fisher–Rao geometry of regular parametric models, we show that, under a high-resolution LAN-scale regime, SMML partitions are asymptotically the pullback, through the maximum-likelihood estimator, of weighted Fisher–Rao Voronoi tessellations in parameter space, with assertion probabilities appearing as additive weights. For regular canonical exponential families, SMML codepoints satisfy a moment-matching condition and admit an interpretation as KL/Bregman centroids, while exact SMML cells are pullbacks of convex polyhedra in sufficient-statistic space. Together, these results show that SMML induces a natural information-geometric quantisation linking entropy-based coding, KL projection, and divergence-based Voronoi geometry. Full article

Scientific flood season segmentation serves as the foundation for determining the flood-limited operating water levels across different periods, providing crucial support for reservoir flood control safety operations and optimal water resource utilization. Under the background of climate change, the traditional static flood-limited water level management model based on fixed dates struggles to adapt to variations in flood season patterns. This study aims to establish a scientifically sound flood season segmentation scheme, providing a basis for dynamic control of flood-limited water levels across different periods, thereby improving water resource utilization efficiency while ensuring flood control safety. This study focuses on the Wuluwati Reservoir and employs the circular distribution method and the Fisher optimal partition method to conduct its flood season segmentation calculations. First, the circular distribution method is used to analyse the concentration and periodic characteristics of flood occurrences in the basin. Subsequently, the Fisher optimal partition method is applied to perform statistical segmentation of the historical hydrological series. Based on this analysis, the flood season of the Wuluwati Reservoir is comprehensively determined as: the pre-flood season from 1 June to 2 July, the main flood season from 3 July to 27 August, and the post-flood season from 28 August to 30 September. To objectively evaluate the rationality of the segmentation results, the improved Cunderlik method was employed to examine the rationality of 15 segmentation schemes based on relative superiority degree. The results show that the scheme with the main flood season from 3 July to 23 August achieves the highest relative superiority degree (0.930). The comprehensively determined segmentation of this study (3 July–27 August) encompasses this optimal interval, demonstrating that the flood season segmentation for the Wuluwati Reservoir is reasonable and effective. Full article

Optimal sensor placement (OSP) for structural health monitoring of large-scale space grid structures must enable reliable identification of localized member deterioration with sparse instrumentation. Modal-based OSP criteria optimize observability of a healthy model but do not directly minimize uncertainty in substructure-level damage parameters. We partition the structure into substructures, simulate axial and biaxial bending stiffness-loss cases, and compute truncated modal flexibility. Each element is encoded by stacked end-node flexibility submatrices over $m=6$ modes. A multi-task, zero-anchored multi-layer perceptron is trained to regress three nonnegative damage parameters and classify damage presence using losses tailored for small-damage accuracy. Sensor sensitivities are obtained by automatic differentiation of the surrogate with respect to flexibility features and aggregated with scenario weights emphasizing critical bending and neighbor-substructure interference scenarios. A greedy D-optimal design then maximizes the log-determinant of a regularized Fisher information matrix under practical coverage constraints; substructure selections are merged into a globally feasible layout. On a representative space grid, the method improves task-oriented identifiability over EFI and MKE across budgets $K_{\mathrm{tot}}=30$–60 (higher-damage D-optimality, lower A-optimality trace, and reduced proxy variance indicators), while yielding lower modal log-determinants. These findings indicate risk-weighted, substructure-first task design as an alternative to purely modal criteria for substructure-level damage-parameter identification. Full article

This study addresses the problem of correcting systematic errors in classical cascade object detectors under severe data scarcity and distribution shift. We focus on the widely used Viola–Jones framework enhanced with a modified Census transform and propose a modular “corrector” architecture that can be attached to an existing detector without retraining it. The key idea is to exploit the blessing of dimensionality: high-dimensional feature vectors constructed from multiple cascade stages are transformed by PCA and whitening into a space where simple linear Fisher discriminants can reliably separate rare error patterns from normal operation using only a few labeled examples. This study presents a novel algorithm designed to correct the outputs of object detectors constructed using the Viola–Jones framework enhanced with a modified census transform. The proposed method introduces several improvements addressing error correction and robustness in data-limited conditions. The approach involves image partitioning through a sliding window of fixed aspect ratio and a modified census transform in which pixel intensity is compared to the mean value within a rectangular neighborhood. Training samples for false negative and false positive correctors are selected using dual Intersection-over-Union (IoU) thresholds and probabilistic sampling of true positive and true negative fragments. Corrector models are trained based on the principles of high-dimensional separability within the paradigm of one- and few-shot learning, utilizing features derived from cascade stages of the detector. Decision boundaries are optimized using Fisher’s rule, with adaptive thresholding to guarantee zero false acceptance. Experimental results indicate that the proposed correction scheme enhances object detection accuracy by effectively compensating for classifier errors, particularly under conditions of scarce training data. On two railway image datasets with only about one thousand images each, the proposed correctors increase Precision from 0.36 to 0.65 on identifier detection while maintaining high Recall (0.98 → 0.94), and improve digit detection Recall from 0.94 to 0.98 with negligible loss in Precision (0.92 → 0.91). These results demonstrate that even under scarce training data, high-dimensional feature separation enables effective one-/few-shot error correction for cascade detectors with minimal computational overhead. Full article

Diverse learning algorithms, optimization methods, and natural selection share a common mathematical structure despite their apparent differences. Here, I show that a simple notational partitioning of change by the Price equation reveals a universal force–metric–bias (FMB) law: $\mathbf{\Delta }\mathit{\theta }=\mathbf{M}\mathbf{f}+\mathbf{b}+\mathit{\xi }$. The force $\mathbf{f}$ drives improvement in parameters, $\mathbf{\Delta }\mathit{\theta }$, in proportion to the slope of performance with respect to the parameters. The metric $\mathbf{M}$ rescales movement by inverse curvature. The bias $\mathbf{b}$ adds momentum or changes in the frame of reference. The noise $\mathit{\xi }$ enables exploration. This framework unifies natural selection, Bayesian updating, Newton’s method, stochastic gradient descent, stochastic Langevin dynamics, Adam optimization, and most other algorithms as special cases of the same underlying process. The Price equation also reveals why Fisher information, Kullback–Leibler divergence, and d’Alembert’s principle arise naturally in learning dynamics. By exposing this common structure, the FMB law provides a principled foundation for understanding, comparing, and designing learning algorithms across disciplines. Full article

The failure of traditional drought indices to capture the dynamic supply–demand imbalance in socio-hydrological systems hinders proactive water management and necessitates novel assessment frameworks. The reservoir drought warning water level, serving as a dynamic threshold indicating supply–demand imbalance, provides a critical basis for drought early warning. From a socio-hydrological drought perspective, this study develops a framework for determining staged and graded soft partition thresholds for reservoir drought warning water levels, encompassing three key stages: water stress analysis, phase classification, and threshold determination. First, water demands for the ecological, agricultural, and domestic sectors were quantified based on hydrological analysis and official operational rules. Second, an optimized KPCA-Fisher model delineated the intra-annual supply–demand dynamics into distinct periods. Thirdly, the soft partition thresholds were formulated by coupling these multi-sectoral demands with water deficit rates using a triangular membership function. Applied to the Xianan Reservoir, the framework yielded distinct drought warning thresholds for the identified main flood, critical demand, and dry seasons. Validation against historical droughts (2019 and 2022) confirmed that these soft thresholds more accurately tracked the drought evolution process compared to traditional hard partitions. Furthermore, a sensitivity analysis identified the ecological water demand methodology as a key factor influencing the thresholds, particularly during the critical demand period. The proposed framework for determining staged and graded reservoir drought warning water levels better reflects the complexity of socio-hydrological systems and provides a scientific basis for refined reservoir drought early warnings and management under changing environments. Full article

The traditional k-means algorithm is widely used in large-scale data clustering because of its easy implementation and efficient process, but it also suffers from the disadvantages of local optimality and poor robustness. In this study, a Csk-means algorithm based on contour similarity is proposed to overcome the drawbacks of the traditional k-means algorithm. For the traditional k-means algorithm, which results in local optimality due to the influence of outliers or noisy data and random selection of the initial clustering centers, the Csk-means algorithm overcomes both drawbacks by combining data lattice transformation and dissimilar interpolation. In particular, the Csk-means algorithm employs Fisher optimal partitioning of the similarity vectors between samples for the process of determining the number of clusters. To improve the robustness of the k-means algorithm to the shape of the clusters, the Csk-means algorithm utilizes contour similarity to compute the similarity between samples during the clustering process. Experimental results show that the Csk-means algorithm provides better clustering results than the traditional k-means algorithm and other comparative algorithms. Full article

Setting the staged flood limit water level (FLWL) through flood season staging is an important means of fully utilizing reservoir flood resources. The widely-used Fisher optimal partition method requires a certain time domain as the basic unit in determining the optimal staging of a flood season. Currently, 5 and 10 days matching the month and solar terms are usually used as the time unit. This study aimed to analyze the influence of other time-domain units (7 and 15 days) that meet the relevant requirements on the staging results and to provide a scientific basis for the selection of time-domain units in flood season staging. The rationality of the staging scheme was tested using the improved Cunderlik method, and the influence of specific basic units in the Fisher optimal partition method on the staging results was evaluated. The highest relative superiority of 0.9876 was found for 5 d, indicating that this is a suitable time-domain unit. The optimal staging result was determined as 20 June for the first segmentation point and 20 August for the second. A comparison of the staged FLWL with a single fixed FLWL showed that the water level was raised by 1.56 m in the pre-flood season, 0.65 m in the main flood season, and 1.37 m in the post-flood season. Water storage increased by 12.79 million m³ during the flood season, effectively alleviating the mismatch between water supply and storage. Full article

In estimating logistic regression models, convergence of the maximization algorithm is critical; however, this may fail. Numerous bias correction methods for maximum likelihood estimates of parameters have been conducted for cases of complete data sets, and also for longitudinal models. Balanced data sets yield consistent estimates from conditional logit estimators for binary response panel data models. When faced with a missing covariates problem, researchers adopt various imputation techniques to complete the data and without loss of generality; consistent estimates still suffice asymptotically. For maximum likelihood estimates of the parameters for logistic regression in cases of imputed covariates, the optimal choice of an imputation technique that yields the best estimates with minimum variance is still elusive. This paper aims to examine the behaviour of the Hessian matrix with optimal values of the imputed covariates vector, which will make the Newton–Raphson algorithm converge faster through a reduced absolute value of the product of the score function and the inverse fisher information component. We focus on a method used to modify the conditional likelihood function through the partitioning of the covariate matrix. We also confirm that the positive moduli of the Hessian for conditional estimators are sufficient for the concavity of the log-likelihood function, resulting in optimum parameter estimates. An increased Hessian modulus ensures the faster convergence of the parameter estimates. Simulation results reveal that model-based imputations perform better than classical imputation techniques, yielding estimates with smaller bias and higher precision for the conditional maximum likelihood estimation of nonlinear panel models. Full article

The traditional flood season division method is cumbersome. In order to make the flood season division elaborate, the Mann–Kendall and cumulative sum of rank difference (CSD) methods were used to detect the abrupt change year of precipitation (p) over the study area from 1969 to 2015. The year of change was determined to be 1995. Taking the 1995 year as a demarcation point of the data, the discriminant model and Fisher optimal partition method were applied for flood division, and a comparison of the results from the two approaches were compared. The discriminant model was found to perform slightly better than the Fisher approach. It was found that abrupt rainfall change has a certain influence on flood season division. The main flood season in the Zhangjia Zhuang reservoir during 1969–2015 was 16 days longer than during 1996–2015, but three days shorter than between 1969–1995. For the Zhangjia Zhuang Reservoir, the flood water level limit can increase up to 2 m according to the results of the flood season division and designed rainfall after abrupt change; in addition, the water storage capacity is 469 million m³ more than that of the traditional reservoir operation mode. Full article

This study analyzes the duration and division of the flood season in the Fenhe River Basin over the period of 1957–2014 based on daily precipitation data collected from 14 meteorological stations. The Mann–Kendall detection, the multiscale moving t-test, and the Fisher optimal partition methods are used to evaluate the impact of climate change on flood season duration and division. The results show that the duration of the flood season has extended in 1975–2014 compared to that in 1957–1974. Specifically, the onset date of the flood season has advanced 15 days, whereas the retreat date of the flood season remains almost the same. The flood season of the Fenhe River Basin can be divided into three stages, and the variations in the onset and retreat dates of each stage are also examined. Corresponding measures are also proposed to better utilize the flood resources to adapt to the flood season variations. Full article