DISK: Differentiable Sparse Kernel Complex for Efficient Spatially-Variant Convolution

TL;DR

DISK introduces a differentiable sparse kernel decomposition method for efficient, high-fidelity spatially-variant convolution suitable for resource-limited devices and real-time applications.

Contribution

It presents a novel differentiable framework for representing complex kernels with sparse samples, enabling efficient spatially-variant filtering without retraining.

Findings

Achieves higher fidelity than simulated annealing.

Significantly lower computational cost than low-rank decompositions.

Effective for mobile imaging and real-time rendering.

Abstract

Image convolution with complex kernels is a fundamental operation in photography, scientific imaging, and animation effects, yet direct dense convolution is computationally prohibitive on resource-limited devices. Existing approximations, such as simulated annealing or low-rank decompositions, either lack efficiency or fail to capture non-convex kernels. We introduce a differentiable kernel decomposition framework that represents a target spatially-variant, dense, complex kernel using a set of sparse kernel samples. Our approach features (i) a decomposition that enables differentiable optimization of sparse kernels, (ii) a dedicated initialization strategy for non-convex shapes to avoid poor local minima, and (iii) a kernel-space interpolation scheme that extends single-kernel filtering to spatially varying filtering without retraining and additional runtime overhead. Experiments on Gaussian and non-convex kernels show that our method achieves higher fidelity than simulated annealing and significantly lower cost than low-rank decompositions. Our approach provides a practical solution for mobile imaging and real-time rendering, while remaining fully differentiable for integration into broader learning pipelines.