Viet-Anh on Software Logo

What is: Dimension-wise Fusion?

SourceDiCENet: Dimension-wise Convolutions for Efficient Networks
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

Dimension-wise Fusion is an image model block that attempts to capture global information by combining features globally. It is an alternative to point-wise convolution. A point-wise convolutional layer applies DD point-wise kernels k_pR3D×1×1\mathbf{k}\_p \in \mathbb{R}^{3D \times 1 \times 1} and performs 3D2HW3D^2HW operations to combine dimension-wise representations of YDimR3D×H×W\mathbf{Y_{Dim}} \in \mathbb{R}^{3D \times H \times W} and produce an output YRD×H×W\mathbf{Y} \in \mathbb{R}^{D \times H \times W}. This is computationally expensive. Dimension-wise fusion is an alternative that can allow us to combine representations of Y_Dim\mathbf{Y\_{Dim}} efficiently. As illustrated in the Figure to the right, it factorizes the point-wise convolution in two steps: (1) local fusion and (2) global fusion.