Mask-Pyramid Network: A Novel Panoptic Segmentation Method.

In this paper, we introduce a novel panoptic segmentation method called the Mask-Pyramid Network. Existing Mask RCNN-based methods first generate a large number of box proposals and then filter them at each feature level, which requires a lot of computational resources, while most of the box proposals are suppressed and discarded in the Non-Maximum Suppression process. Additionally, for panoptic segmentation, it is a problem to properly fuse the semantic segmentation results with the Mask RCNN-produced instance segmentation results. To address these issues, we propose a new mask pyramid mechanism to distinguish objects and generate much fewer proposals by referring to existing segmented masks, so as to reduce computing resource consumption. The Mask-Pyramid Network generates object proposals and predicts masks from larger to smaller sizes. It records the pixel area occupied by the larger object masks, and then only generates proposals on the unoccupied areas. Each object mask is represented as a H × W × 1 logit, which fits well in format with the semantic segmentation logits. By applying SoftMax to the concatenated semantic and instance segmentation logits, it is easy and natural to fuse both segmentation results. We empirically demonstrate that the proposed Mask-Pyramid Network achieves comparable accuracy performance on the Cityscapes and COCO datasets. Furthermore, we demonstrate the computational efficiency of the proposed method and obtain competitive results.

Unveiling Enhanced Electrostatic Repulsion in Silica Nanosphere Assembly: Formation Dynamics of Body-Centered-Cubic Colloidal Crystals.

We demonstrate the effective establishment of long-range electrostatic interactions among colloidal silica nanospheres through acid treatment, enabling their assembly into colloidal crystals at remarkably low concentrations. This novel method overcomes the conventional limitation in colloidal silica assembly by removing entrapped NH4+ ions and enhancing the electrical double layer (EDL) thickness, offering a time-efficient alternative to increase electrostatic interactions compared with methods like dialysis. The increased EDL thickness facilitates the assembly of SiO2 nanospheres into a body-centered-cubic lattice structure at low particle concentrations, allowing for broad spectrum tunability and high tolerance to particle size polydispersity. Further, we uncover a disorder-order transition during colloidal crystallization at low particle concentrations, with the optimal concentration for crystal formation governed by both thermodynamic and kinetic factors. This work not only provides insights into assembly mechanisms but also paves the way for the design and functionalization of colloidal silica-based photonic crystals in diverse applications.

Laser Induced Coffee-Ring Structure through Solid-Liquid Transition for Color Printing.

Metallic micro/nano structures with special physicochemical properties have undergone rapid development owing to their broad applications in micromachines and microdevices. Ultrafast laser processing is generally accepted as an effective technology for functional structures manufacture, however, the controllable fabrication of specific metallic micro/nano structures remains a challenge. Here, this work proposes a novel strategy of laser induced transient solid-liquid transition to fabricate unique structures. Through modulating the transient state of metal from solid to liquid phase using the initial pulse excitation, the subsequent ultrafast pulse-induced recoil pressure can suppress the plasma emission and removal of liquid phase metals, resulting in the controllable fabrication of coffee-ring structures. The solid-liquid transition dynamics, which related with the transient reflectivity and plasma intensity, are revealed by established two temperature model coupled with molecular dynamics model. The coffee-ring structure exhibits tunable structure color owing to various optical response, which can be used for color printing with large scale and high resolution. This work provides a promising strategy for fabricating functional micro/nano structures, which can greatly broaden the potential applications.

Dynamical Mechanism Analysis of Three Neuroregulatory Strategies on the Modulation of Seizures.

This paper attempts to explore and compare the regulatory mechanisms of optogenetic stimulation (OS), deep brain stimulation (DBS) and electromagnetic induction on epilepsy. Based on the Wilson-Cowan model, we first demonstrate that the external input received by excitatory and inhibitory neural populations can induce rich dynamic bifurcation behaviors such as Hopf bifurcation, and make the system exhibit epileptic and normal states. Then, both OS and DBS are shown to be effective in controlling the epileptic state to a normal low-level state, and the stimulus parameters have a broad effective range. However, electromagnetic induction cannot directly control epilepsy to this desired state, even if it can significantly reduce the oscillation frequency of neural populations. One main difference worth noting is that the high spatiotemporal specificity of OS allows it to target inhibitory neuronal populations, whereas DBS and electromagnetic induction can only stimulate excitatory as well as inhibitory neuronal populations together. Next, the propagation behavior of epilepsy is explored under a typical three-node feedback loop structure. An increase in coupling strength accelerates and exacerbates epileptic activity in other brain regions. Finally, OS and DBS applied to the epileptic focus play similar positive roles in controlling the behavior of the area of seizure propagation, while electromagnetic induction still only achieves unsatisfactory effects. It is hoped that these dynamical results can provide insights into the treatment of epilepsy as well as other neurological disorders.

Edge-Sensitive Left Ventricle Segmentation Using Deep Reinforcement Learning.

Deep reinforcement learning (DRL) has been utilized in numerous computer vision tasks, such as object detection, autonomous driving, etc. However, relatively few DRL methods have been proposed in the area of image segmentation, particularly in left ventricle segmentation. Reinforcement learning-based methods in earlier works often rely on learning proper thresholds to perform segmentation, and the segmentation results are inaccurate due to the sensitivity of the threshold. To tackle this problem, a novel DRL agent is designed to imitate the human process to perform LV segmentation. For this purpose, we formulate the segmentation problem as a Markov decision process and innovatively optimize it through DRL. The proposed DRL agent consists of two neural networks, i.e., First-P-Net and Next-P-Net. The First-P-Net locates the initial edge point, and the Next-P-Net locates the remaining edge points successively and ultimately obtains a closed segmentation result. The experimental results show that the proposed model has outperformed the previous reinforcement learning methods and achieved comparable performances compared with deep learning baselines on two widely used LV endocardium segmentation datasets, namely Automated Cardiac Diagnosis Challenge (ACDC) 2017 dataset, and Sunnybrook 2009 dataset. Moreover, the proposed model achieves higher F-measure accuracy compared with deep learning methods when training with a very limited number of samples.

Seismic sedimentology response characteristics of nearshore terminal fan in the suning area of Raoyang Sag, Bohai Bay Basin, NE China.

A nearshore terminal fan is a special water system formed in arid environments. The characterisation of its thin-channel sand bodies has long been a challenge restricting oil and gas exploration. This study takes the Suning area of the Raoyang Sag as an example and uses the principles of seismic sedimentology to conduct seismic sedimentary research on the nearshore terminal fan of the first member of the Palaeogene Shahejie Formation (Es1) based on three-dimensional seismic, logging, and core analysis. Seven fourth-order sequences (SQV7) were identified within Es1, deposited by a fluvial river system terminating at the contracting bank of a lake. Prograding terminal fan sedimentary facies on a gentle slope zone were observed in the root mean square seismic attributes after spectral decomposition. We have successfully resolved the sandstone within the studied terminal fan system using a 90° phase conversion of the seismic data and red-green-blue (RGB) fusion of the various seismic attributes. The upper subsegment of the Shahejie Formation developed extensive nearshore terminal fan sedimentation, and the seismic sedimentological response characteristics were mainly channel-like and strip-shaped geomorphic systems deposited on gentle slope zones, indicating distributary channels and distal basin sedimentation. This study enriches our understanding of nearshore fans and provides ideas for predicting favourable sand bodies in this type of sedimentary facies.

Deep learning algorithm-enabled sediment characterization techniques to determination of water saturation for tight gas carbonate reservoirs in Bohai Bay Basin, China.

Understanding water saturation levels in tight gas carbonate reservoirs is vital for optimizing hydrocarbon production and mitigating challenges such as reduced permeability due to water saturation (Sw) and pore throat blockages, given its critical role in managing capillary pressure in water drive mechanisms reservoirs. Traditional sediment characterization methods such as core analysis, are often costly, invasive, and lack comprehensive spatial information. In recent years, several classical machine learning models have been developed to address these shortcomings. Traditional machine learning methods utilized in reservoir characterization encounter various challenges, including the ability to capture intricate relationships, potential overfitting, and handling extensive, multi-dimensional datasets. Moreover, these methods often face difficulties in dealing with temporal dependencies and subtle patterns within geological formations, particularly evident in heterogeneous carbonate reservoirs. Consequently, despite technological advancements, enhancing the reliability, interpretability, and applicability of predictive models remains imperative for effectively characterizing tight gas carbonate reservoirs. This study employs a novel data-driven strategy to prediction of water saturation in tight gas reservoir powered by three recurrent neural network type deep/shallow learning algorithms-Gated Recurrent Unit (GRU), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Support Vector Machine (SVM), K-nearest neighbor (KNN) and Decision tree (DT)-customized to accurately forecast sequential sedimentary structure data. These models, optimized using Adam's optimizer algorithm, demonstrated impressive performance in predicting water saturation levels using conventional petrophysical data. Particularly, the GRU model stood out, achieving remarkable accuracy (an R-squared value of 0.9973) with minimal errors (RMSE of 0.0198) compared to LSTM, RNN, SVM, KNN and, DT algorithms, thus showcasing its proficiency in processing extensive datasets and effectively identifying patterns. By achieving unprecedented accuracy levels, this study not only enhances the understanding of sediment properties and fluid saturation dynamics but also offers practical implications for reservoir management and hydrocarbon exploration in complex geological settings. These insights pave the way for more reliable and efficient decision-making processes, thereby advancing the forefront of reservoir engineering and petroleum geoscience.

A Machine Learning-Combined Flexible Sensor for Tactile Detection and Voice Recognition.

Intelligent sensors have attracted substantial attention for various applications, including wearable electronics, artificial intelligence, healthcare monitoring, and human-machine interactions. However, there still remains a critical challenge in developing a multifunctional sensing system for complex signal detection and analysis in practical applications. Here, we develop a machine learning-combined flexible sensor for real-time tactile sensing and voice recognition through laser-induced graphitization. The intelligent sensor with a triboelectric layer can convert local pressure to an electrical signal through a contact electrification effect without external bias, which has a characteristic response behavior when exposed to various mechanical stimuli. With the special patterning design, a smart human-machine interaction controlling system composed of a digital arrayed touch panel is constructed to control electronic devices. Based on machine learning, the real-time monitoring and recognition of the changes of voice are achieved with high accuracy. The machine learning-empowered flexible sensor provides a promising platform for the development of flexible tactile sensing, real-time health detection, human-machine interaction, and intelligent wearable devices.

HSGAN: Hyperspectral Reconstruction From RGB Images With Generative Adversarial Network.

Hyperspectral (HS) reconstruction from RGB images denotes the recovery of whole-scene HS information, which has attracted much attention recently. State-of-the-art approaches often adopt convolutional neural networks to learn the mapping for HS reconstruction from RGB images. However, they often do not achieve high HS reconstruction performance across different scenes consistently. In addition, their performance in recovering HS images from clean and real-world noisy RGB images is not consistent. To improve the HS reconstruction accuracy and robustness across different scenes and from different input images, we present an effective HSGAN framework with a two-stage adversarial training strategy. The generator is a four-level top-down architecture that extracts and combines features on multiple scales. To generalize well to real-world noisy images, we further propose a spatial-spectral attention block (SSAB) to learn both spatial-wise and channel-wise relations. We conduct the HS reconstruction experiments from both clean and real-world noisy RGB images on five well-known HS datasets. The results demonstrate that HSGAN achieves superior performance to existing methods. Please visit https://github.com/zhaoyuzhi/HSGAN to try our codes.

SVCNet: Scribble-Based Video Colorization Network With Temporal Aggregation.

In this paper, we propose a scribble-based video colorization network with temporal aggregation called SVCNet. It can colorize monochrome videos based on different user-given color scribbles. It addresses three common issues in the scribble-based video colorization area: colorization vividness, temporal consistency, and color bleeding. To improve the colorization quality and strengthen the temporal consistency, we adopt two sequential sub-networks in SVCNet for precise colorization and temporal smoothing, respectively. The first stage includes a pyramid feature encoder to incorporate color scribbles with a grayscale frame, and a semantic feature encoder to extract semantics. The second stage finetunes the output from the first stage by aggregating the information of neighboring colorized frames (as short-range connections) and the first colorized frame (as a long-range connection). To alleviate the color bleeding artifacts, we learn video colorization and segmentation simultaneously. Furthermore, we set the majority of operations on a fixed small image resolution and use a Super-resolution Module at the tail of SVCNet to recover original sizes. It allows the SVCNet to fit different image resolutions at the inference. Finally, we evaluate the proposed SVCNet on DAVIS and Videvo benchmarks. The experimental results demonstrate that SVCNet produces both higher-quality and more temporally consistent videos than other well-known video colorization approaches. The codes and models can be found at https://github.com/zhaoyuzhi/SVCNet.

Three-dimensional patterning of MoS2 with ultrafast laser.

Transition metal chalcogenides, a special two-dimensional (2D) material emerged in recent years, possess unique optoelectronic properties and have been used to fabricate various optoelectronic devices. While it is essential to manufacture multifunctional devices with complex nanostructures for practical applications, 2D material devices present a tendency toward miniaturization. However, the controllable fabrication of complex nanostructures on 2D materials remains a challenge. Herein, we propose a method to create designed three-dimensional (3D) patterns on the MoS2 surface by modulating the interaction between an ultrafast laser and MoS2. Three different nanostructures, including flat, bulge, and craters, can be fabricated through laser-induced surface morphology transformation, which is related to thermal diffusion, oxidation, and ablation processes. The MoS2 field effect transistor is fabricated by ultrafast laser excitation which exhibits enhanced electrical properties. This study provides a promising strategy for 3D pattern fabrication, which is helpful for the development of multifunctional microdevices.

ShaTure: Shape and Texture Deformation for Human Pose and Attribute Transfer.

In this paper, we present a novel end-to-end pose transfer framework to transform a source person image to an arbitrary pose with controllable attributes. Due to the spatial misalignment caused by occlusions and multi-viewpoints, maintaining high-quality shape and texture appearance is still a challenging problem for pose-guided person image synthesis. Without considering the deformation of shape and texture, existing solutions on controllable pose transfer still cannot generate high-fidelity texture for the target image. To solve this problem, we design a new image reconstruction decoder - ShaTure which formulates shape and texture in a braiding manner. It can interchange discriminative features in both feature-level space and pixel-level space so that the shape and texture can be mutually fine-tuned. In addition, we develop a new bottleneck module - Adaptive Style Selector (AdaSS) Module which can enhance the multi-scale feature extraction capability by self-recalibration of the feature map through channel-wise attention. Both quantitative and qualitative results show that the proposed framework has superiority compared with the state-of-the-art human pose and attribute transfer methods. Detailed ablation studies report the effectiveness of each contribution, which proves the robustness and efficacy of the proposed framework.

Transition dynamics and optogenetic controls of generalized periodic epileptiform discharges.

This paper aims to analyze possible mechanisms underlying the generation of generalized periodic epileptiform discharges (GPEDs), especially to design targeted optogenetic regulation strategies. First and foremost, inspired by existing physiological experiments, we propose a new computational framework by introducing a second inhibitory neuronal population and related synaptic connections into the classic Liley mean field model. The improved model can simulate the basic normal and abnormal brain activities mentioned in previous studies, but much to our relief, it perfectly reproduces some types of GPEDs that match the clinical records. Specifically, results show that disinhibitory synaptic connections between inhibitory interneuronal populations are closely related to the occurrence, transition and termination of GPEDs, including delaying the occurrence of GPEDs caused by the excitatory AMPAergic autapses and regulating the transition process of GPEDs bidirectionally, which support the conjecture that selective changes of synaptic connections can trigger GPEDs. Additionally, we creatively offer six optogenetic strategies with dual targets. They can all control GPEDs well, just as experiments reveal that optogenetic stimulation of inhibitory interneurons can suppress abnormal activities in epilepsy or other brain diseases. More importantly, 1:1 coordinated reset stimulation with one period rest is concluded as the optimal strategy after taking into account the energy consumption and control effect. Hope these results provide feasible references for pathophysiological mechanisms of GPEDs.

Study on the relationship of leisure sports tourism with the health of the elderly / Estudo sobre a relação do turismo esportivo de lazer com a saúde do idoso / Estudio sobre la relación de turismo deportivo de ocio con la salud del anciano

ABSTRACT Objective: To explore the influence of leisure sports tourism activities on the health of elderly tourists. Methods: Through investigating the leisure sports tourism activities of some elderly tourists who participated in travel agencies, the health status and quality of life of elderly tourists who participated in leisure sports tourism activities for a long time (exercising group) were compared with those who did not or occasionally participated in physical exercise (non-exercising group), to understand the influence of leisure sports tourism activities on the health of elderly tourists. Results: Regarding the attitude of participating in sports leisure tourism activities, the survey results show that the elderly tourists reached a basic consensus about participating in sports leisure tourism activities. Different types of landscapes have significant differences in improving tourists' anxiety. The proportion of "very satisfied" in the exercise group was significantly higher than that in the non-exercise group. The duration of the trip has a significant impact on improving tourists' anxiety. Conclusion: We should strengthen the advertising of leisure sport tourism activities, let more elderly tourists participate in leisure sport tourism activities, strengthen the behavioral guidance of elderly tourists in leisure sport tourism activities, and help them improve their health. Level of evidence II; Therapeutic studies - investigation of treatment results.

RESUMO Objetivo: Explorar a influência das atividades de turismo esportivo de lazer sobre a saúde dos turistas idosos. Método: Através da investigação das atividades de turismo de lazer esportivo de alguns turistas idosos que participaram de agências de viagem, o estado de saúde e qualidade de vida dos turistas idosos que participaram de atividades de turismo de lazer esportivo por um longo tempo (grupo de exercícios) foram comparados com aqueles que não participaram ou ocasionalmente participaram de exercícios físicos (grupo não praticante), para entender a influência das atividades de turismo de lazer esportivo sobre a saúde dos turistas idosos. Resultados: Quanto à atitude de participação em atividades de turismo de lazer esportivo, os resultados da pesquisa mostram que os turistas idosos chegaram a um consenso básico sobre a participação em atividades de turismo de lazer esportivo. Diferentes tipos de paisagens têm diferenças significativas para melhorar a ansiedade dos turistas. A proporção de "muito satisfeitos" no grupo de exercícios foi significativamente maior do que a do grupo de não-exercícios. A duração da viagem tem um impacto significativo na melhoria da ansiedade dos turistas. Conclusão: Devemos fortalecer a propaganda das atividades de turismo esportivo de lazer, deixar que mais turistas idosos participem das atividades de turismo esportivo de lazer, fortalecer a orientação do comportamento dos turistas idosos em atividades de turismo esportivo de lazer, e ajudá-los a melhorar sua saúde. Nível de evidência II; Estudos terapêuticos - investigação de resultados de tratamento.

RESUMEN Objetivo: Explorar la influencia de las actividades de turismo deportivo de ocio en la salud de los turistas de edad avanzada. Método: Mediante la investigación de las actividades de turismo deportivo de ocio de algunos turistas de la tercera edad que participaron en agencias de viajes, se comparó el estado de salud y la calidad de vida de los turistas de la tercera edad que participaron en actividades de turismo deportivo de ocio durante un largo periodo (grupo ejercitante) con los que no hicieron ejercicio físico o lo hicieron ocasionalmente (grupo no ejercitante), para comprender la influencia de las actividades de turismo deportivo de ocio en la salud de los turistas de la tercera edad. Resultados: En cuanto a la actitud de participación en actividades de turismo de ocio deportivo, los resultados de la investigación muestran que los turistas de edad avanzada llegaron a un consenso básico sobre la participación en actividades de turismo de ocio deportivo. Los distintos tipos de paisajes presentan diferencias significativas a la hora de mejorar la ansiedad de los turistas. La proporción de "muy satisfechos" en el grupo de ejercicio fue significativamente mayor que en el grupo de no ejercicio. La duración del viaje tiene un impacto significativo en la mejora de la ansiedad de los turistas. Conclusión: Debemos reforzar la publicidad de las actividades de turismo deportivo de ocio, dejar que más turistas de edad avanzada participen en las actividades de turismo deportivo de ocio, reforzar la orientación del comportamiento de los turistas de edad avanzada en las actividades de turismo deportivo de ocio y ayudarles a mejorar su salud. Nivel de evidencia II; Estudios terapéuticos - investigación de resultados de tratamiento.