所以本文提出了一种新的时间序列预测损失函数,称为Patch-wise Structural (PS) 损失,旨在通过引入局部统计特性来提高时间序列预测的准确性。 为了说明 MSE 的局限性,我们来看图 1 中三个具有相似 MSE 值但预测质量不同的预测结果。图 1a 中的预测与真实值的相关性较差,未能捕捉到整体的方向和模式。相比之下,图 1b 中的预测虽然与真实
More importantly, this paper inject a new fusion mechanism into the conventional Laplacian pyramid fusion framework, namely the non-standardized structural patch decomposition algorithm. While retaining its original advantages, it reduces the computational complexity of the algorithm. Experimental results on ...
While video synthesis de- mands global coherence across full-frame contexts, super- resolution prioritizes localized detail restoration atop exist- ing structural content. Specifically, we augment the base model with two con- ditional branches: a patch bra...
Particularly, utilizing lamination parameters and a patch-wise lay-up approach allowed further enhanced structural performance beyond those offered by conventional laminated composites. In the first level, lamination parameters were utilized as design variables of the lay-up optimization to solve the ...
PPformer, for low-light image enhancement. PPformer is a CNN-transformer hybrid network that is divided into three parts: local-branch, global-branch, and Dual Cross-Attention. Each part plays a vital role in PPformer. Specifically, the local-branch extracts local structural information using a...
时间序列预测在交通、天气和金融等多个领域具有重要应用价值。然而,现有的预测模型大多依赖于点对点的损失函数,如均方误差,这些方法忽略了时间序列数据中的结构性依赖关系,导致难以准确捕捉复杂的时间模式。 所以本文提出了一种新的时间序列预测损失函数,称为Patch-wise Structural (PS) 损失,旨在通过引入局部统计特性来提...
Structural magnetic resonance imaging (MRI) has been widely utilized for analysis and diagnosis of brain diseases. Automatic segmentation of brain tumors is a challenging task for computer-aided diagnosis due to low-tissue contrast in the tumor subregions. To overcome this, we devise a novel pixel...
Pixel-wise segmentationCNN3D to 2D conversionBrain tumorMRIStructural magnetic resonance imaging (MRI) has been widely utilized for analysis and diagnosis of brain diseases. Automatic segmentation of brain tumors is a challenging task for computer-aided diagnosis due to low-tissue contrast in the ...
Specifically, the local-branch extracts local structural information using a stack of Wide Enhancement Modules , and the global-branch provides the refining global information by Cross Patch Module and Global Convolution Module. Besides, different from self-attention, we use extracted global semantic ...