Search

Cloud contamination significantly impairs optical remote sensing images (RSIs), reducing their utility for Earth observation. The traditional cloud removal techniques, often reliant on deep learning, generally aim for holistic image reconstruction, which may inadvertently alter the intrinsic qualities of cloud-free areas, leading to image distortions. To address this issue, we propose a multi-stage frequency attention network (MFCRNet), a progressive paradigm for optical RSI cloud removal. MFCRNet hierarchically deploys frequency cloud removal modules (FCRMs) to refine the cloud edges while preserving the original characteristics of the non-cloud regions in the frequency domain. Specifically, the FCRM begins with a frequency attention block (FAB) that transforms the features into the frequency domain, enhancing the differentiation between cloud-covered and cloud-free regions. Moreover, a non-local attention block (NAB) is employed to augment and disseminate contextual information effectively. Furthermore, we introduce a collaborative loss function that amalgamates semantic, boundary, and frequency-domain information. The experimental results on the RICE1, RICE2, and T-Cloud datasets demonstrate that MFCRNet surpasses the contemporary models, achieving superior performance in terms of mean absolute error (MAE), root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM), validating its efficacy regarding the cloud removal from optical RSIs. [ABSTRACT FROM AUTHOR]

The generation of co-speech gestures for digital humans is an emerging area in the field of virtual human creation. Prior research has progressed by using acoustic and semantic information as input and adopting a classification method to identify the person's ID and emotion for driving co-speech gesture generation. However, this endeavor still faces significant challenges. These challenges go beyond the intricate interplay among co-speech gestures, speech acoustic, and semantics; they also encompass the complexities associated with personality, emotion, and other obscure but important factors. This paper introduces "DiT-Gestures", a speech-conditional diffusion-based and non-autoregressive transformer-based generative model with the WavLM pre-trained model and a dynamic mask attention network (DMAN). It can produce individual and stylized full-body co-speech gestures by only using raw speech audio, eliminating the need for complex multimodal processing and manual annotation. Firstly, considering that speech audio contains acoustic and semantic features and conveys personality traits, emotions, and more subtle information related to accompanying gestures, we pioneer the adaptation of WavLM, a large-scale pre-trained model, to extract the style from raw audio information. Secondly, we replace the causal mask by introducing a learnable dynamic mask for better local modeling in the neighborhood of the target frames. Extensive subjective evaluation experiments are conducted on the Trinity, ZEGGS, and BEAT datasets to confirm WavLM's and the model's ability to synthesize natural co-speech gestures with various styles. [ABSTRACT FROM AUTHOR]

As one of the important internal curing agents for cement-based materials, inorganic/organic composite water absorbing materials, which has hydrophilic groups and three-dimensional network structures, exhibits excellent water absorption, water retention, and water release properties. By introducing inorganic/organic composite water absorbing materials into cement-based materials, their water absorption and release characteristics can automatically adjust the internal relative humidity of cement-based materials, promote hydration reactions, reduce self shrinkage and early cracking of cement-based materials, and thereby improve mechanical properties and durability. The preparation methods, structural characteristics and absorption/desorption behavior of inorganic/organic composite internal curing materials (CICA) were introduced. The impact on the hydration process, microstructure, macroscopic properties and durability of cement-based materials were summarized. The development prospects of CICA in concrete applications were prospected, providing theoretical guidance and technical reference for the application of CICA in cement-based internal curing. [ABSTRACT FROM AUTHOR]

Semantic segmentation plays a pivotal role in the intelligent interpretation of remote sensing images (RSIs). However, conventional methods predominantly focus on learning representations within the spatial domain, often resulting in suboptimal discriminative capabilities. Given the intrinsic spectral characteristics of RSIs, it becomes imperative to enhance the discriminative potential of these representations by integrating spectral context alongside spatial information. In this paper, we introduce the spectrum-space collaborative network (SSCNet), which is designed to capture both spectral and spatial dependencies, thereby elevating the quality of semantic segmentation in RSIs. Our innovative approach features a joint spectral–spatial attention module (JSSA) that concurrently employs spectral attention (SpeA) and spatial attention (SpaA). Instead of feature-level aggregation, we propose the fusion of attention maps to gather spectral and spatial contexts from their respective branches. Within SpeA, we calculate the position-wise spectral similarity using the complex spectral Euclidean distance (CSED) of the real and imaginary components of projected feature maps in the frequency domain. To comprehensively calculate both spectral and spatial losses, we introduce edge loss, Dice loss, and cross-entropy loss, subsequently merging them with appropriate weighting. Extensive experiments on the ISPRS Potsdam and LoveDA datasets underscore SSCNet's superior performance compared with several state-of-the-art methods. Furthermore, an ablation study confirms the efficacy of SpeA. [ABSTRACT FROM AUTHOR]

The service life of herbicide-resistant varieties can be prolonged by identification of intercrossing-resistance of existing herbicide-resistant rapeseed and the screening of application of other safe herbicide in dosage forms. In this study7 a new germplasm PN19 and transgenic plants of Arabidopsis and tobacco resistant to SU herbicides were used as materials. Five classes of ALS inhibitory herbicides with five concentration gradients were sprayed at 3-5 leaf stage of rapeseed. The intercrossing-resistance of PN19 to five herbicides was determined by resistance identification. Results showed that plant intercrossing-resistance against SU and SCT class was the strongest' followed by IMI and TP class7 whereas that against PB class was the lowest. Evaluation of herbicide intercrossing-resistance in transgenic plants and analysis of in vitro activity of tobacco ALS enzymes revealed this resistance function to various herbicides was resulted from the decreased sensitivity of target ALS enzymes to herbicides. In addition' the expression of PN19 herbicide-resistance gene after treatment with five herbicides was higher than that of wild type. The study provides an opportunity by mixing use of ALS herbicides to integratedly control weeds in the field. [ABSTRACT FROM AUTHOR]

It can better guide the trend of network public opinion and improve the relationship between doctors and patients by analyzing the sentiment evolutionary trends of doctor-patient relationships in recent years and assisting the government in providing the theoretical basis for decision-making and intervention. In the large-scale public opinion text of doctors and patients in the complex context of the Internet, it has constructed a sentiment classifier in the downstream tasks of the pre-training model BERT. It is combined with LDA topic extraction technology for multi-dimensional emotional evolution analysis. Finally, the ARIMA model is combined to predict sentiment trends. The experimental results show that the sentiment prediction model accuracy based on LDA-BERT reaches 98%. The average prediction error of the ARIMA time series prediction of the doctor-patient public opinion model is less than 11.25%. It is demonstrated that the algorithm presented in this research can be effectively applied to the multi-dimensional monitoring and analysis of large-scale doctor-patient public opinion evolution. [ABSTRACT FROM AUTHOR]

Increasing the efficiency of resource allocation is the basis and guarantee for boosting high-quality economic development. Based on the panel data of Chinese industrial enterprises and cities from 2008 to 2013, this paper studies the influence of infrastructure construction demand shocks represented by local government debt expansion on the efficiency of sectoral resource allocation from the perspective of sectoral linkage. According to the empirical findings, local government debt significantly reduces the resource allocation efficiency of manufacturing sector that is highly related to infrastructure construction. This conclusion is still tenable after the robustness test using the simulated local government debt as an instrumental variable. Further mechanism tests show that there are two reasons for the decline of the efficiency of resource allocation in manufacturing sector that is highly related to infrastructure construction. First, more product demands and investments brought by the expansion of local government debt flow to less productive enterprises in the sector. Second, resource misallocation reduces the probability of high-productivity enterprises entering the market and low-productivity enterprises exiting the market, and the effect is more prominent in cities with high dependence on state-owned enterprises and high pressure on officials to be promoted. According to this study, the performance management of local government debt should be further strengthened, and particular attention should be paid to the influence of local government debt on enterprise investment and financing crowding out and resource misallocation. [ABSTRACT FROM AUTHOR]