Generative models have attracted considerable attention for speech separation tasks, and among these, diffusion-based methods are being explored. Despite the notable success of diffusion techniques in generation tasks, their adaptation to speech separation has encountered challenges, notably slow convergence and suboptimal separation outcomes. To address these issues and enhance the efficacy of diffusion-based speech separation, we introduce EDSep, a novel single-channel method grounded in score matching via stochastic differential equation (SDE). This method enhances generative modeling for speech source separation by optimizing training and sampling efficiency. Specifically, a novel denoiser function is proposed to approximate data distributions, which obtains ideal denoiser outputs. Additionally, a stochastic sampler is carefully designed to resolve the reverse SDE during the sampling process, gradually separating speech from mixtures. Extensive experiments on databases such as WSJ0-2mix, LRS2-2mix, and VoxCeleb2-2mix demonstrate our proposed method’s superior performance over existing diffusion and discriminative models, validating its efficacy.

Intelligent reflecting surface (IRS) can significantly increase wireless transmission rates by intelligently adjusting reflection coefficients, i.e., phase shifts and amplitudes, and optimizing the deployment position. However, most existing works mainly focus on the optimization of IRSs’ reflection coefficients but have not deeply delved into the issue of IRS deployment, especially for the scenarios with multiple cooperative IRSs. Therefore, we propose to jointly optimize IRSs’ locations together with reflection coefficients and mobile users’ (MUs’) rate allocations for double-IRS-aided communication systems, where one base station (BS) transmits data to two groups of MUs through the cooperation of one passive IRS (PIRS) and one active simultaneous transmitting and reflecting IRS (ASTAR-IRS). Unlike PIRS, the ASTAR-IRS can simultaneously transmit and reflect signals, and then it can achieve full-space coverage and can be more flexibly deployed. Moreover, for further improving communication rates, the BS adopts the rate-splitting multiple access (RSMA) technique to transmit data. First, considering the statistics of channel state information (CSI), we formulate a rate maximization problem to maximize the expectations of all MUs’ aggregate data rates. Then, based on the deep reinforcement learning, we develop a joint optimization scheme to optimize the cooperative IRSs’ positions as well as reflection coefficients and MUs’ rate allocations for the case with perfect CSI. Third, we extend our work to the more realistic case with imperfect CSI. Finally, we verify and evaluate the performances of our proposed schemes through extensive simulations, which show that MUs’ sum data rates can increase by about 21.4% and 51.4% by using RSMA and optimizing the two cooperative IRSs’ locations, respectively.

The fraction of mulched film is of great significance for evaluating the agricultural water-saving effect and controlling environmental plastic pollution. Unfortunately, there is no work has been done to obtain this parameter due to the mixed pixel issue of satellite imagery with medium and low resolutions, till now. In this study, we proposed a novel two-step framework for mapping fraction of mulched film based on very high resolution satellite observation and deep learning. The first step is extracting the extent of the mulched film based on a new few-shot learning model named PT-CNN (Parameter Transition Convolutional Neural Network), which aims to increase the extraction accuracy and overcome the lack of labeled training data. The second step is retrieving the fraction of the mulched film at pixel scale based on a spectrum analysis method. The result shows that the proposed PT-CNN outperforms several state-of-the-art methods in mulched film extraction, with F1-scores at 97.09% and 98.65% for white and black mulched film, respectively. Meanwhile, the retrieved pixel scale fraction of mulched film shows high consistency to in-situ measurement, with a MAE of 0.0321. The proposed method can be useful in agricultural water resource management and environmental governance.

Unsupervised domain adaptation(UDA) aims to mitigate the performance drop of models tested on the target domain, due to the domain shift from the target to sources. Most UDA segmentation methods focus on the scenario of solely single source domain. However, in practical situations data with gold standard could be available from multiple sources (domains), and the multi-source training data could provide more information for knowledge transfer. How to utilize them to achieve better domain adaptation yet remains to be further explored. This work investigates multi-source UDA and proposes a new framework for medical image segmentation. Firstly, we employ a multi-level adversarial learning scheme to adapt features at different levels between each of the source domains and the target, to improve the segmentation performance. Then, we propose a multi-model consistency loss to transfer the learned multi-source knowledge to the target domain simultaneously. Finally, we validated the proposed framework on two applications, i.e., multi-modality cardiac segmentation and cross-modality liver segmentation. The results showed our method delivered promising performance and compared favorably to state-of-the-art approaches.

Accurate canopy and single-tree mapping is important to obtain information on the ecological structure and biogeophysical parameters for forests. Although some airborne radars can retrieve canopy and single-tree information within a smaller area, the optical satellite imagery-based approaches for rapidly and accurately mapping them over a large region are still limited. In this study, based onEucalyptuscanopy and single-tree texture and spectral features, we proposed a mapping approach using the combinations of image morphology, the Otsu method, and an adaptive iterative erosion algorithm (EUMAP). Then, we applied the commonly used red/green/blue bands from the high-resolution satellite images, which are freely available, to map the canopy and single-tree inEucalyptusplantations in southern China. EUMAP consists of two steps: (i)Eucalyptuscanopy identification for various canopy density regions; (ii) adaptive iterative erosion to separate single-tree. Our study was conducted in the Chengzhong and Liubei districts of Liuzhou city, China. The accuracy evaluation was carried out in the state-owned Sanmenjiang Forest Farm. The results showed that the average F1 score for mapping canopy and single-tree reached 88.34% and 86.40%, respectively. For the whole study area, there were 7033021Eucalyptustrees and the average density was 819 trees per hectare. The approach adopted in this study, combining the prior knowledges about image morphology and single-tree texture features ofEucalyptusplantations, was highly efficient for satellite image processing and had excellent applicability to large-scaleEucalyptusplantations mapping. Our study highlights the necessary of prior knowledges for forest mapping using satellite images without requiring a training sample and provides a universal approach of accurately large-scale mapping for specific forest species with common red/green/blue images.