CoordGate: Efficiently Computing Spatially-Varying Convolutions in Convolutional Neural Networks
This addresses the challenge of handling spatially-varying blur in optical imaging for computer vision applications, offering a more robust solution, though it appears incremental as it builds on existing CNN and deblurring techniques.
The paper tackles the problem of efficiently computing spatially-varying convolutions in CNNs for image deblurring, proposing CoordGate, a lightweight module that uses a multiplicative gate and coordinate encoding to enable selective filter amplification based on spatial position, and it outperforms conventional approaches in experiments.
Optical imaging systems are inherently limited in their resolution due to the point spread function (PSF), which applies a static, yet spatially-varying, convolution to the image. This degradation can be addressed via Convolutional Neural Networks (CNNs), particularly through deblurring techniques. However, current solutions face certain limitations in efficiently computing spatially-varying convolutions. In this paper we propose CoordGate, a novel lightweight module that uses a multiplicative gate and a coordinate encoding network to enable efficient computation of spatially-varying convolutions in CNNs. CoordGate allows for selective amplification or attenuation of filters based on their spatial position, effectively acting like a locally connected neural network. The effectiveness of the CoordGate solution is demonstrated within the context of U-Nets and applied to the challenging problem of image deblurring. The experimental results show that CoordGate outperforms conventional approaches, offering a more robust and spatially aware solution for CNNs in various computer vision applications.