Can I read TVConv: Efficient Translation Variant Convolution for Layout-aware Visual Processing on EtoBox?

TVConv: Efficient Translation Variant Convolution for Layout-aware Visual Processing by Chen, Jierun; He, Tianlang; Zhuo, Weipeng; Ma, Li; Ha, Sangtae; Chan, S. -H. Gary is a scholarly article available to read on EtoBox.

What is TVConv: Efficient Translation Variant Convolution for Layout-aware Visual Processing about?

As convolution has empowered many smart applications, dynamic convolution further equips it with the ability to adapt to diverse inputs. However, the static and dynamic convolutions are either layout-agnostic or computation-heavy, making it inappropriate for layout-specific applications, e.g., face recognition and medical image segmentation. We observe that these applications naturally exhibit the characteristics of large intra-image (spatial) variance and small cross-image variance. This observation motivates our efficient translation variant convolution (TVConv) for layout-aware visual processing. Technically, TVConv is composed of affinity maps and a weight-generating block. While affinity maps depict pixel-paired relationships gracefully, the weight-generating block can be explicitly overparameterized for better training while maintaining efficient inference. Although conceptually simple, TVConv significantly improves the efficiency of the convolution and can be readily plugged into various network architectures. Extensive experiments on face recognition show that TVConv reduces the computational cost by up to 3.1x and improves the corresponding throughput by 2.3x while maintai

Author

Chen, Jierun; He, Tianlang; Zhuo, Weipeng; Ma, Li; Ha, Sangtae; Chan, S. -H. Gary

Published

2022

Language

EN