Scale Steerable Filters for Locally Scale-Invariant Convolutional Neural Networks

06/10/2019
by   Rohan Ghosh, et al.
0

Augmenting transformation knowledge onto a convolutional neural network's weights has often yielded significant improvements in performance. For rotational transformation augmentation, an important element to recent approaches has been the use of a steerable basis i.e. the circular harmonics. Here, we propose a scale-steerable filter basis for the locally scale-invariant CNN, denoted as log-radial harmonics. By replacing the kernels in the locally scale-invariant CNN lsi_cnn with scale-steered kernels, significant improvements in performance can be observed on the MNIST-Scale and FMNIST-Scale datasets. Training with a scale-steerable basis results in filters which show meaningful structure, and feature maps demonstrate which demonstrate visibly higher spatial-structure preservation of input. Furthermore, the proposed scale-steerable CNN shows on-par generalization to global affine transformation estimation methods such as Spatial Transformers, in response to test-time data distortions.

READ FULL TEXT
research
08/31/2019

Towards Learning Affine-Invariant Representations via Data-Efficient CNNs

In this paper we propose integrating a priori knowledge into both design...
research
01/16/2019

Extension of Convolutional Neural Network with General Image Processing Kernels

We applied pre-defined kernels also known as filters or masks developed ...
research
11/28/2021

Implicit Equivariance in Convolutional Networks

Convolutional Neural Networks(CNN) are inherently equivariant under tran...
research
11/24/2014

Scale-Invariant Convolutional Neural Networks

Even though convolutional neural networks (CNN) has achieved near-human ...
research
06/10/2018

Transformationally Identical and Invariant Convolutional Neural Networks through Symmetric Element Operators

Mathematically speaking, a transformationally invariant operator, such a...
research
07/31/2018

Scale equivariance in CNNs with vector fields

We study the effect of injecting local scale equivariance into Convoluti...
research
07/30/2018

Transformationally Identical and Invariant Convolutional Neural Networks by Combining Symmetric Operations or Input Vectors

Transformationally invariant processors constructed by transformed input...

Please sign up or login with your details

Forgot password? Click here to reset