Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation

TL;DR

This paper introduces Deformable Kernels, a novel convolutional operator that adapts the effective receptive field directly during runtime to better handle object deformations, improving recognition performance.

Contribution

The work proposes a new method for directly adapting the effective receptive field in convolutional networks, providing a generic, drop-in replacement for rigid kernels that enhances deformation handling.

Findings

Deformable Kernels outperform prior deformation modeling methods.

The approach effectively adapts to object deformations across multiple tasks.

Theoretical analysis confirms ERF is determined by data sampling and kernel values.

Abstract

Convolutional networks are not aware of an object's geometric variations, which leads to inefficient utilization of model and data capacity. To overcome this issue, recent works on deformation modeling seek to spatially reconfigure the data towards a common arrangement such that semantic recognition suffers less from deformation. This is typically done by augmenting static operators with learned free-form sampling grids in the image space, dynamically tuned to the data and task for adapting the receptive field. Yet adapting the receptive field does not quite reach the actual goal -- what really matters to the network is the "effective" receptive field (ERF), which reflects how much each pixel contributes. It is thus natural to design other approaches to adapt the ERF directly during runtime. In this work, we instantiate one possible solution as Deformable Kernels (DKs), a family of novel and generic convolutional operators for handling object deformations by directly adapting the ERF while leaving the receptive field untouched. At the heart of our method is the ability to resample the original kernel space towards recovering the deformation of objects. This approach is justified with theoretical insights that the ERF is strictly determined by data sampling locations and kernel values. We implement DKs as generic drop-in replacements of rigid kernels and conduct a series of empirical studies whose results conform with our theories. Over several tasks and standard base models, our approach compares favorably against prior works that adapt during runtime. In addition, further experiments suggest a working mechanism orthogonal and complementary to previous works.