Second, to avoid excessive information loss caused by a deep network, we designed a three-level feature fusion network to supplement as much original information as possible into the output feature map. Third, we have introduced a transformer module in the last layer of the...
Zhang, Y., Han, S., Zhang, Z., et al.: CF-GAN: cross-domain feature fusion generative adversarial network for text-to-image synthesis. Visual Comput. 39(4), 1283–1293 (2022) Google Scholar Peng, D., Yang, W., Liu, C., et al.: SAM-GAN: self-attention supporting multi-stag...
Then, a cross-stage based feature fusion model was designed to replace the skip concatenation of original UNet, which was composed of Wide Range Attention unit, Small-kernel Local Attention unit, and Inverted Bottleneck unit. WRA was employed to capture global attention, whose large convolution ...
Following that, the feature pyramid network (FPN) is connected in series with a bottom-up feature pyramid structure to realize the feature fusion. The results of model Ablation experiment and baggage scanning image detection show that the cascade cross-stage YOLOv3 model significantly improves the ...
The paper introduces a robust multi-feature cross-fusion approach, i.e., a multi-feature dual-stage cross manifold attention network, namely, MF-DCMANet, which essentially relies on the complementary information between different features to enhance the representation ability of targets. In...
Multi-resolution feature fusionCurriculum learningSurveillance systemsFace recognition for surveillance remains a complex challenge due to the disparity between low-resolution (LR) face images captured by surveillance cameras and the typically high-resolution (HR) face images in databases. To address this ...
FEATURE extractionIn remote sensing image fusion, the conventional Convolutional Neural Networks (CNNs) extract local features of the image through layered convolution, which is limited by the receptive field and struggles to capture global features. Transformer utilizes self-attention to capture long-...
The proposed method consists of a feature-wised fusion block using the attention mechanism and the strength estimation block using FFT and sequential representations(FTB). We first describe the speech enhancement problem mathematically, after which we compared the proposed method with some well-known ...
A dual-stage feature cross-fusion representation framework is proposed, respectively named Cross-Feature Network (CFN) and Cross-Manifold Attention (CMA); In MF-DCMANet, handcrafted monogenic features and polarization features are combined with deep features to improve target recognition accuracy; By ...