Hyperspectral imaging (HSI) has been widely used in agricultural applications for non-destructive estimation of plant nutrient composition and precise quantification of sample nutritional elements. Recently, 3D reconstruction methods, such as Neural Radiance Field (NeRF), have been used to create implicit neural representations of HSI scenes. This capability enables the rendering of hyperspectral channel compositions at every spatial location, thereby helping localize the target object's nutrient composition both spatially and spectrally. However, it faces limitations in training time and rendering speed. In this paper, we propose Diffusion-Denoised Hyperspectral Gaussian Splatting (DD-HGS), which enhances the state-of-the-art 3D Gaussian Splatting (3DGS) method with wavelength-aware spherical harmonics, a Kullback-Leibler divergence-based spectral loss, and a diffusion-based denoiser to enable 3D explicit reconstruction of the hyperspectral scenes for the entire spectral range. We present extensive evaluations on diverse real-world hyperspectral scenes from the Hyper-NeRF dataset to show the effectiveness of our DD-HGS. The results demonstrate that DD-HGS achieves the new state-of-the-art performance compared to all the previously published methods. Project page: https://dragonpg2000.github.io/DDHGS-website/

With the continuous advancement of wireless communication technologies and the growing diversity of modulation schemes, Automatic Modulation Recognition (AMR) plays a crucial role in spectrum sensing and signal analysis. However, traditional deep learning models with fixed architectures often face challenges in simultaneously achieving high-precision classification of known modulation types and effectively rejecting unknown modulated signals. To address this, we propose a novel framework combining a custom neural network with OpenMax. First, we introduce a time-frequency diagram channel alongside the I/Q channels to capture both time-domain waveform details and spectral structure, which enhances the model's ability to distinguish between complex modulation schemes and detect unknown signals in low-SNR environments. The network input consists of the time-frequency diagram, I component, and Q component. Moreover, the OpenMax open-set recognition model is incorporated to perform tail probability fitting and re-calibration of the neural network's activation vectors, enabling precise classification of known categories and dynamic rejection of unknown signals. Extensive simulations under challenging conditions—including Rayleigh fading, carrier frequency offset, and AWGN with SNR ranging from ‑10 dB to 20 dB—demonstrate that the proposed method achieves a closed-set classification accuracy of 92.21% and an unknown signal detection rate of 84.38%. Furthermore, computational complexity and runtime analysis show that the proposed framework achieves end-to-end inference latencies of approximately 2.1 ms on GPU and 3.7 ms on CPU, including time-frequency preprocessing, confirming its practical feasibility for real-time deployment. Compared to conventional methods like dual-channel CNN, dual-channel/triple-channel ResNet, and CNN+LSTM, the proposed approach demonstrates quite good performance in both closed-set and open-set tasks, effectively validating the effectiveness of the time-frequency fusion strategy and uncertainty calibration mechanism in improving model robustness and generalization.

The solution to boundary value problems is of great significance in industrial software applications. In this paper, we propose a novel deep learning method for simulating stress field distributions in simply supported beams, aiming to serve as a solver for stress boundary value problems. Our regression network, Stress-EA, utilizes the convolution encoder module and additive attention to accurately estimate the stress in the beam. By comparing the Stress-EA prediction results with the stress values calculated using ABAQUS, we achieve a mean absolute error (MAE) of less than 0.06. This indicates a high level of consistency between the stress values obtained from the two approaches. Moreover, the prediction time of Stress-EA is significantly shorter, taking only 0.0011s, compared to the calculation time of ABAQUS, which is 16.91s. This demonstrates the high accuracy and low computational latency of our model. Furthermore, our model exhibits smaller model parameters, requires less computation, and has a shorter prediction time compared to training results obtained using classic and advanced networks. To accelerate training, we utilize data parallel methods, achieving up to 1.89 speedup on a dual-GPU platform without compromising accuracy. This advancement enhances the computing efficiency for large-scale industrial software applications.

We propose a novel hybrid cable-based robot with manipulator and camera for high-accuracy, medium-throughput plant monitoring in a vertical hydroponic farm and, as an example application, demonstrate non-destructive plant mass estimation. Plant monitoring with high temporal and spatial resolution is important to both farmers and researchers to detect anomalies and develop predictive models for plant growth. The availability of high-quality, off-the-shelf structure-from-motion (SfM) and photogrammetry packages has enabled a vibrant community of roboticists to apply computer vision for non-destructive plant monitoring. While existing approaches tend to focus on either high-throughput (e.g. satellite, unmanned aerial vehicle (UAV), vehicle-mounted, conveyor-belt imagery) or high-accuracy/robustness to occlusions (e.g. turn-table scanner or robot arm), we propose a middle-ground that achieves high accuracy with a medium-throughput, highly automated robot. Our design pairs the workspace scalability of a cable-driven parallel robot (CDPR) with the dexterity of a 4 degree-of-freedom (DoF) robot arm to autonomously image many plants from a variety of viewpoints. We describe our robot design and demonstrate it experimentally by collecting daily photographs of 54 plants from 64 viewpoints each. We show that our approach can produce scientifically useful measurements, operate fully autonomously after initial calibration, and produce better reconstructions and plant property estimates than those of over-canopy methods (e.g. UAV). As example applications, we show that our system can successfully estimate plant mass with a Mean Absolute Error (MAE) of 0.586g and, when used to perform hypothesis testing on the relationship between mass and age, produces p-values comparable to ground-truth data (p=0.0020 and p=0.0016, respectively).

Abstract Recently self-supervised learning is gaining popularity for Graph Neural Networks (GNN) by leveraging unlabeled data. Augmentation plays a key role in self-supervision. While there is a common set of image augmentation methods that preserve image labels in general, graph augmentation methods do not guarantee consistent graph semantics and are usually domain dependent. Existing self-supervised GNN models often handpick a small set of augmentation techniques that limit the performance of the model. In this paper, we propose a common set of graph augmentation methods to a wide range of GNN tasks, and rely on the Pareto optimality to select and balance among these possibly conflicting augmented versions, called Pareto Graph Contrastive Learning (PGCL) framework. We show that while random selection of the same set of augmentation leads to slow convergence or even divergence, PGCL converges much faster with lower error rate. Extensive experiments on multiple datasets of different domains and scales demonstrate superior or comparable performance of PGCL.

BACKGROUND: The human leukocyte antigen (HLA) gene family plays a key role in the immune response and thus is crucial in many biomedical and clinical settings. Utilizing Sanger sequencing, the golden standard technology for HLA typing enables accurate identification of HLA alleles in high-resolution. However, only the commercial software, such as uTYPE, SBT-Assign, and SBTEngine, and very few open-source tools could be applied to perform HLA typing based on Sanger sequencing. RESULTS: We developed a user-friendly, cross-platform and open-source desktop application, known as SOAPTyping, for Sanger-based typing in HLA class I and II alleles. SOAPTyping can produce accurate results with a comprehensible protocol and featured functions. Moreover, SOAPTyping supports a more advanced group-specific sequencing primers (GSSP) module to solve the ambiguous typing results. We used SOAPTyping to analyze 36 samples with known HLA typing from the University of California Los Angeles (UCLA) International HLA DNA Exchange platform and 100 anonymous clinical samples, and the HLA typing results from SOAPTyping are identical to the golden results and 5.5 times faster than commercial software uTYPE, which shows the usability of SOAPTyping. CONCLUSIONS: We introduce the SOAPTyping as the first open-source and cross-platform HLA typing software with the capability of producing high-resolution HLA typing predictions from Sanger sequence data.