Bidirectional Cascade Research Articles

FoSp: Focus and Separation Network for Early Smoke Segmentation

Early smoke segmentation (ESS) enables the accurate identification of smoke sources, facilitating the prompt extinguishing of fires and preventing large-scale gas leaks. But ESS poses greater challenges than conventional object and regular smoke segmentation due to its small scale and transparent appearance, which can result in high miss detection rate and low precision. To address these issues, a Focus and Separation Network (FoSp) is proposed. We first introduce a Focus module employing bidirectional cascade which guides low-resolution and high-resolution features towards mid-resolution to locate and determine the scope of smoke, reducing the miss detection rate. Next, we propose a Separation module that separates smoke images into a pure smoke foreground and a smoke-free background, enhancing the contrast between smoke and background fundamentally, improving segmentation precision. Finally, a Domain Fusion module is developed to integrate the distinctive features of the two modules which can balance recall and precision to achieve high F_beta. Futhermore, to promote the development of ESS, we introduce a high-quality real-world dataset called SmokeSeg, which contains more small and transparent smoke than the existing datasets. Experimental results show that our model achieves the best performance on three available smoke segmentation datasets: SYN70K (mIoU: 83.00%), SMOKE5K (F_beta: 81.6%) and SmokeSeg (F_beta: 72.05%). The code can be found at https://github.com/LujianYao/FoSp.

Large-scale self-organization in dry turbulent atmospheres.

How turbulent convective fluctuations organize to form larger-scale structures in planetary atmospheres remains a question that eludes quantitative answers. The assumption that this process is the result of an inverse cascade was suggested half a century ago in two-dimensional fluids, but its applicability to atmospheric and oceanic flows remains heavily debated, hampering our understanding of the energy balance in planetary systems. We show using direct numerical simulations with spatial resolutions of 122882 × 384 points that rotating and stratified flows can support a bidirectional cascade of energy, in three dimensions, with a ratio of Rossby to Froude numbers comparable to that of Earth's atmosphere. Our results establish that, in dry atmospheres, spontaneous order can arise through an inverse cascade to the largest spatial scales.

PlantBiCNet: A new paradigm in plant science with bi-directional cascade neural network for detection and counting

Deep learning is increasingly popular for precise plant detection and counting in computer vision applications. Despite the rising cross-disciplinary research in this field, an important issue has been neglected: due to the inherent constraints imposed by the natural growth patterns of plants, the costly and time-consuming process of collecting plant image data, resulting in a scarcity of real-world available data. Consequently, researchers are often constrained to exploring large-scale data representations or limited applications. Currently, still no generalized network architecture excels in demanding environments. To address this, we introduce two new bi-directional cascade neural network paradigms, PlantBiCNet and its lightweight variant, PlantBiCNet-Lite. Unlike previous research, we place greater emphasis on the network decoding process itself rather than the encoder. Our bi-directional cascade and weight fusion decoding approach maximizes high-level semantics and low-level spatial data, enhancing the localization signal in a way that respects gradient diversity. We evaluate PlantBiCNet and PlantBiCNet-Lite on five public plant detection and counting datasets and a novel plant dataset. The average precision AP50 on these six small-scale datasets reached 0.905, with a high coefficient of determination R2 of 0.935, while the reported frame rate was 196 FPS. These results significantly outperform state-of-the-art computer vision methods. Moreover, we found that the introduction of the Dropout mechanism markedly enhances the model's performance and generalization ability. Further analysis supports our experiments and provides valuable insights into network optimization. Overall, our organization's open datasets and network modeling details provide a solid foundation for future research in the field of plant science.

ДВОНАПРАВЛЕНИЙ КАСКАДНИЙ ПЕРЕТВОРЮВАЧ ПОСТІЙНОЇ НАПРУГИ ДЛЯ ПОТУЖНОЇ СИСТЕМИ ЕНЕРГОНАКОПИЧЕННЯ В МЕРЕЖАХ ЕЛЕКТРОПОСТАЧАННЯ SMART GRID

Electromagnetic processes in a bidirectional cascade DC converter with the possibility of increasing and decreasing the output voltage levels in both directions of energy transfer for electric power storage systems are considered. Using the method of averaging in the state space, a mathematical model of the pulse converter was obtained. Analytical ex-pressions for calculating the ripple characteristics of a bidirectional cascade converter have been determined, with the help of which you can calculate the main parameters of the constant voltage converter, which provide permissible volt-age ripples on the DC buses and power choke current. To illustrate the developed calculation methodology, the ripple characteristics and parameters of a 30 kW converter for various load modes with 400V and 700V DC busbars for use in energy storage systems are determined. The bidirectional converter was calculated, analyzed and simulated in Math-cad/PSim packages at idle speed and at different load levels in order to confirm the effectiveness of the conducted re-search. Ref. 7, fig. 2. Keywords: method of averaging in state space, electric power storage systems, bidirectional constant voltage converter, non-isolated four-quadrant bidirectional converter.

Automatic Identification of Thaw Slumps Based on Neural Network Methods and Thaw Slumping Susceptibility

Thaw slumping is a periglacial process that occurs on slopes in cold environments, where the ground becomes unstable and the surface slides downhill due to saturation with water during thawing. In this study, GaoFen-1 remote sensing and fused multi-source feature data were used to automatically map thaw slumping landforms in the Beilu River Basin of the Qinghai–Tibet Plateau. The bi-directional cascade network structure was used to extract edges at different scales, where an individual layer was supervised by labeled edges at its specific scale, rather than directly applying the same supervision to all convolutional neural network outputs. Additionally, we conducted a 5-year multi-scale feature analysis of small baseline subset interferometric synthetic aperture radar deformation, normalized difference vegetation index, and slope, among other features. Our study analyzed the performance and accuracy of three methods based on edge object supervised learning and three preconfigured neural networks, ResNet101, VGG16, and ResNet152. Through verification using site surveys and multi-data fusion results, we obtained the best ResNet101 model score of intersection over union of 0.85 (overall accuracy of 84.59%).The value of intersection over union of the VGG and ResNet152 are 0.569 and 0.773, respectively. This work provides a new insight for the potential feasibility of applying the designed edge detection method to map diverse thaw slumping landforms in larger areas with high-resolution images.

Helical fluid and (Hall)-MHD turbulence: a brief review.

Helicity, a measure of the breakage of reflectional symmetry representing the topology of turbulent flows, contributes in a crucial way to their dynamics and to their fundamental statistical properties. We review several of their main features, both new and old, such as the discovery of bi-directional cascades or the role of helical vortices in the enhancement of large-scale magnetic fields in the dynamo problem. The dynamical contribution in magnetohydrodynamic of the cross-correlation between velocity and induction is discussed as well. We consider next how turbulent transport is affected by helical constraints, in particular in the context of magnetic reconnection and fusion plasmas under one- and two-fluid approximations. Central issues on how to construct turbulence models for non-reflectionally symmetric helical flows are reviewed, including in the presence of shear, and we finally briefly mention the possible role of helicity in the development of strongly localized quasi-singular structures at small scale.This article is part of the theme issue ‘Scaling the turbulence edifice (part 2)’.

A relic sketch extraction framework based on detail-aware hierarchical deep network

As the first step of the restoration process of painted relics, sketch extraction plays an important role in cultural research. However, sketch extraction suffers from serious disease corrosion, which results in broken lines and noise. To overcome these problems, we propose a deep learning-based hierarchical sketch extraction framework for painted cultural relics. We design the sketch extraction process into two stages: coarse extraction and fine extraction. In the coarse extraction stage, we develop a novel detail-aware bi-directional cascade network that integrates flow-based difference-of-Gaussians (FDoG) edge detection and a bi-directional cascade network (BDCN) under a transfer learning framework. It not only uses the pre-trained strategy to extenuate the requirements of large datasets for deep network training but also guides the network to learn the detail characteristics by the prior knowledge from FDoG. For the fine extraction stage, we design a new multiscale U-Net (MSU-Net) to effectively remove disease noise and refine the sketch. Specifically, all the features extracted from multiple intermediate layers in the decoder of MSU-Net are fused for sketch predication. Experimental results showed that the proposed method outperforms the other seven state-of-the-art methods in terms of visual and quantitative metrics and can also deal with complex backgrounds.

BDCN: Bi-Directional Cascade Network for Perceptual Edge Detection.

Exploiting multi-scale representations is critical to improve edge detection for objects at different scales. To extract edges at dramatically different scales, we propose a bi-directional cascade network (BDCN) architecture, where an individual layer is supervised by labeled edges at its specific scale, rather than directly applying the same supervision to different layers. Furthermore, to enrich multi-scale representations learned by each layer of BDCN, we introduce a scale enhancement module (SEM), which utilizes dilated convolution to generate multi-scale features, instead of using deeper CNNs. These new approaches encourage the learning of multi-scale representations in different layers and detect edges that are well delineated by their scales. Learning scale dedicated layers also results in a compact network with a fraction of parameters. We evaluate our method on three datasets, i.e., BSDS500, NYUDv2, and Multicue, and achieve ODS F-measure of 0.832, 2.7 percent higher than current state-of-the-art on the BSDS500 dataset. We also applied our edge detection result to other vision tasks. Experimental results show that, our method further boosts the performance of image segmentation, optical flow estimation, and object proposal generation.

Helicity Dynamics, Inverse, and Bidirectional Cascades in Fluid and Magnetohydrodynamic Turbulence: A Brief Review

We briefly review helicity dynamics, inverse and bidirectional cascades in fluid and magnetohydrodynamic (MHD) turbulence, with an emphasis on the latter. The energy of a turbulent system, an invariant in the nondissipative case, is transferred to small scales through nonlinear mode coupling. Fifty years ago, it was realized that, for a two‐dimensional fluid, energy cascades instead to larger scales and so does magnetic excitation in MHD. However, evidence obtained recently indicates that, in fact, for a range of governing parameters, there are systems for which their ideal invariants can be transferred, with constant fluxes, to both the large scales and the small scales, as for MHD or rotating stratified flows, in the latter case including quasi‐geostrophic forcing. Such bidirectional, split, cascades directly affect the rate at which mixing and dissipation occur in these flows in which nonlinear eddies interact with fast waves with anisotropic dispersion laws, due, for example, to imposed rotation, stratification, or uniform magnetic fields. The directions of cascades can be obtained in some cases through the use of phenomenological arguments, one of which we derive here following classical lines in the case of the inverse magnetic helicity cascade in electron MHD. With more highly resolved data sets stemming from large laboratory experiments, high‐performance computing, and in situ satellite observations, machine learning tools are bringing novel perspectives to turbulence research. Such algorithms help devise new explicit subgrid‐scale parameterizations, which in turn may lead to enhanced physical insight, including in the future in the case of these new bidirectional cascades.

Critical transitions in thin layer turbulence

We investigate a model of thin layer turbulence that follows the evolution of the two-dimensional motions $\boldsymbol{u}_{2D}(x,y)$ along the horizontal directions $(x,y)$ coupled to a single Fourier mode along the vertical direction ($z$) of the form $\boldsymbol{u}_{q}(x,y,z)=[v_{x}(x,y)\sin (qz),v_{y}(x,y)\sin (qz),v_{z}(x,y)\cos (qz)]$, reducing thus the system to two coupled, two-dimensional equations. The model, despite its simplicity and ad hoc construction, displays a rich behaviour. Its reduced dimensionality allows a thorough investigation of the transition from a forward to an inverse cascade of energy as the thickness of the layer $H=\unicode[STIX]{x03C0}/q$ is varied. Starting from a thick layer and reducing its thickness it is shown that two critical heights are met: (i) one for which the forward unidirectional cascade (similar to three-dimensional turbulence) transitions to a bidirectional cascade transferring energy to both small and large scales and (ii) one for which the bidirectional cascade transitions to a unidirectional inverse cascade when the layer becomes very thin (similar to two-dimensional turbulence). The two critical heights are shown to have different properties close to criticality that we are able to analyse with numerical simulations for a wide range of Reynolds numbers and aspect ratios.

A Common Model for the Approximate Analysis of Tandem Queueing Systems With Blocking

This work proposes a common model that unifies several approximate methods for the analysis of tandem queueing systems with blocking. Two abstract functions are introduced to characterize the relationship between starvation and blockage in neighboring machines. Using the functions, the equilibrium equations are determined, and a general procedure is developed to solve them. The sufficient conditions for the existence and uniqueness of a solution are obtained. More generally, the proposed model leads to the concept of a bidirectional cascade system, in which each subsystem has both forward and backward inputs and forward and backward outputs. Finally, the theory of the common model is applied to specific systems to demonstrate its applicability.

Formation of the bi-directional energy cascade in low-frequency damped wave-turbulent systems

We demonstrate that in wave-turbulent systems, an inverse energy flux can be formed in addition to the conventional direct energy flux if large-scale wave damping is present. Such a bi-directional energy cascade can be understood in terms of the wave energy transfer on the temperature wave background. The formation of the bi-directional cascade provides an effective mechanism for global coupling between the scales where the turbulent fluctuations at small scales are affected by the wave distribution at large scales due to the total energy conservation. We discuss physical systems where this scenario is realized.

Fluctuation and noise propagation in phenotypic transition cascades of clonal populations.

Quantitative modeling of fluctuations of each phenotype is a crucial step towards a fundamental understanding of noise propagation through various phenotypic transition cascades. The theoretical formulas for noise propagation in various phenotypic transition cascades are derived by using the linear noise approximation of master equation and the logarithmic gain. By virtue of the theoretical formulas, we study the noise propagation in bidirectional and unidirectional phenotypic transition cascades, respectively. It is found that noise propagation in these two phenotypic transition cascades evidently differs: In the bidirectional cascade, a systemic random environment is provided by a correlated global component. The total noise of each phenotype is mainly determined by the intrinsic noise and the transmitted noise from other phenotypes. The intrinsic noise enlarged by interconversion through an added part shows a novel noise propagation mechanism. However, in the unidirectional cascade, the random environment of each downstream phenotype is provided by upstream phenotypes. The total noise of each downstream phenotype is mainly determined by the transmitted noises from upstream phenotypes. The intrinsic noise and the conversion noise can propagate in both bidirectional and unidirectional phenotypic transition cascades.

Bidirectional energy cascade in surface capillary waves.

Based on an experiment and simulations, we report that an energy cascade in surface capillary waves can be bidirectional, that is, can simultaneously flow towards large and small wavelength scales from the pumping scales. The bidirectional energy cascade provides an effective global coupling mechanism between the scales. We show that formation of the bidirectional cascade leads to creation of large-scale, large-amplitude waves on the fluid surface.

Whatever next? Predictive brains, situated agents, and the future of cognitive science

Brains, it has recently been argued, are essentially prediction machines. They are bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions. This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing. Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to adaptive success. This target article critically examines this "hierarchical prediction machine" approach, concluding that it offers the best clue yet to the shape of a unified science of mind and action. Sections 1 and 2 lay out the key elements and implications of the approach. Section 3 explores a variety of pitfalls and challenges, spanning the evidential, the methodological, and the more properly conceptual. The paper ends (sections 4 and 5) by asking how such approaches might impact our more general vision of mind, experience, and agency.

Model for bidirectional transmission in an open cascade of optical amplifiers

We derive a model for signal propagation and the generation and propagation of amplified spontaneous emission (ASE) in a completely open bidirectional cascade of optical amplifiers. The effect of single Rayleigh scattering of both signal and ASE is included. Our model predicts and our experiments substantially verify that the dominant gain saturating mechanism is the Rayleigh backscattered ASE centered around the 1535 nm peak in the ASE spectrum and that the spectral shape of the output ASE becomes much more sharply peaked around the 1535 nm gain peak. In an experimental open cascade of three EDFAs the ASE spectral density at the 1535 nm gain peak is increased by as much as 8-9 dB, resulting in gain compression ranging from 1-2 dB near 1550 nm to as much as 35 dB nearer 1538 nm. This combined with multiple Rayleigh scattering of the ASE around our 1538 and 1540 nm signal wavelengths results in Rayleigh scattering induced reductions of the optical signal to ASE ratio of 1-2 dB at the output of the first amplifier increasing to 36 dB at the output of the third amplifier each signal traverses.