Qiu Guan

Three-dimensional (3D) gravity inversion remains a fundamentally ill-posed and non-unique problem, where reconstructing subsurface density distributions from two-dimensional (2D) surface gravity observations is severely constrained by depth-dependent sensitivity decay and ambiguity in mass placement. Classical Tikhonov-based inversion algorithms partially alleviate these issues through explicit regularization but often suffer from the skin effect, oversmoothed deep anomalies, and substantial computational cost. Recent deep learning (DL) approaches provide a data-driven alternative; however, conventional convolutional architectures still struggle with semantic inconsistency during multi-scale feature fusion, limited depth-awareness, and extreme class imbalance in geophysical models.This paper presents GravityVNet, a physics-aware deep neural framework tailored for high-fidelity 3D gravity inversion. The architecture is built upon a Residual V-Net backbone enhanced with a 3D Coordinate Attention (CoordAtt) mechanism to encode long-range spatial dependencies and explicitly preserve depth-direction information, which helps compensate for depth-dependent attenuation effects commonly associated with the skin effect. To further improve structural recovery, we introduce Multi-Scale Attentional Feature Fusion (AFF) modules, which adaptively recalibrate skip-connected features and bridge semantic gaps between encoder and decoder pathways. A Deep Supervision training strategy, combined with a Hybrid Loss function integrating Focal Tversky Loss and Total Variation (TV) regularization, enforces boundary sharpness, suppresses artifacts, and strengthens convergence stability. Extensive experiments on large-scale, physics-consistent synthetic datasets generated by SimPEG indicate that GravityVNet achieves competitive performance compared to standard baseline methods. The results indicate that the model has improved in terms of geometric fidelity, quantitative accuracy, and physical consistency with the observed gravity response. In addition, the validation of the measured data in the Vinton Salt Dome also proves the effectiveness of the model in solving composite anomalies and deep structures.

Dilated convolution is a powerful technique for expanding the receptive field without increasing the convolution kernel size, making it highly valuable in image segmentation tasks. However, some challenges still exist, such as sparse feature extraction at high dilation rates and limited neighboring information interaction. To address these issues, a novel approach, termed Dynamic Dilated Convolution (DDConv), is proposed for adaptive feature extraction by dynamically adjusting dilation rates in this paper. In DDConv, dilated convolutions are divided into four groups along the channel, each assigned distinct dilation rates. Attention weights are then used to aggregate information across these groups, enabling the convolution kernel to dynamically select the dilation rate during feature map convolution. Experimental results demonstrate the effectiveness of the proposed methods, with DDConv delivering outstanding performance. These findings suggest promising implications for DDNeXt as a novel backbone construction approach for future developments in the field.

In existing medical Region of Interest (ROI) detection, there lacks an algorithm that can simultaneously satisfy both real-time performance and accuracy, not meeting the growing demand for automatic detection in medicine. Although the basic YOLO framework ensures real-time detection due to its fast speed, it still faces challenges in maintaining precision concurrently. To alleviate the above problems, we propose a novel model named Lightweight Shunt Matching-YOLO (LSM-YOLO), with Lightweight Adaptive Extraction (LAE) and Multipath Shunt Feature Matching (MSFM). Firstly, by using LAE to refine feature extraction, the model can obtain more contextual information and high-resolution details from multiscale feature maps, thereby extracting detailed features of ROI in medical images while reducing the influence of noise. Secondly, MSFM is utilized to further refine the fusion of high-level semantic features and low-level visual features, enabling better fusion between ROI features and neighboring features, thereby improving the detection rate for better diagnostic assistance. Experimental results demonstrate that LSM-YOLO achieves 48.6% AP on a private dataset of pancreatic tumors, 65.1% AP on the BCCD blood cell detection public dataset, and 73.0% AP on the Br35h brain tumor detection public dataset. Our model achieves state-of-the-art performance with minimal parameter cost on the above three datasets. The source codes are at: https://github.com/VincentYuuuuuu/LSM-YOLO.

Due to the effective performance of multi-scale feature fusion, Path Aggregation FPN (PAFPN) is widely employed in YOLO detectors. However, it cannot efficiently and adaptively integrate high-level semantic information with low-level spatial information simultaneously. We propose a new model named MAF-YOLO in this paper, which is a novel object detection framework with a versatile neck named Multi-Branch Auxiliary FPN (MAFPN). Within MAFPN, the Superficial Assisted Fusion (SAF) module is designed to combine the output of the backbone with the neck, preserving an optimal level of shallow information to facilitate subsequent learning. Meanwhile, the Advanced Assisted Fusion (AAF) module deeply embedded within the neck conveys a more diverse range of gradient information to the output layer. Furthermore, our proposed Re-parameterized Heterogeneous Efficient Layer Aggregation Network (RepHELAN) module ensures that both the overall model architecture and convolutional design embrace the utilization of heterogeneous large convolution kernels. Therefore, this guarantees the preservation of information related to small targets while simultaneously achieving the multi-scale receptive field. Finally, taking the nano version of MAF-YOLO for example, it can achieve 42.4% AP on COCO with only 3.76M learnable parameters and 10.51G FLOPs, and approximately outperforms YOLOv8n by about 5.1%. The source code of this work is available at: https://github.com/yang-0201/MAF-YOLO.

With the prosperity of deep learning techniques, salient object detection in remote sensing images (RSI-SOD) is concomitantly in full flourishing. However, due to the inherent challenges such as uncertainty in object quantities and scales, cluttered backgrounds, and blurred edges arising from shadows, most current approaches struggle for salient feature learning with the aid of heavy model architecture, yet often result in barely satisfactory performance. Some methods compromise model complexity to improve efficiency, albeit with significantly degraded results. To earn a satisfactory balance of efficacy and efficiency, we propose a new network for RSI-SOD, namely SFANet, based on progressive semantic flow and uncertainty-aware refinement. Specifically, we design a global semantic enhancement block (GSEB) to reduce background interference and accurately localize salient objects of varying quantities and scales, which further consists of three modularized components, i.e., semantic extraction module (SEM), interscale fusion module (IFM), and deep semantic graph-inference module (DSGM). SEM together with IFM contributes to the effective aggregation of multi-scale contexts by extracting fused and progressive semantic cues. DSGM performs semantic inference to better localize salient objects with irregularities in scale and topological structure. Furthermore, we present an uncertainty-aware refinement module (URM) to recognize salient objects in cluttered backgrounds and effectively suppress shadows. Extensive experiments are conducted on three RSI-SOD datasets, from which superior results can be achieved by our SFANet, outperforming the other cutting-edge methods. The code is available at https://github.com/ZhengJianwei2/SFANet.