Exploiting SIFT Descriptor for Rotation Invariant Convolutional Neural Network

TL;DR

This paper introduces a CNN model that replaces max-pooling with SIFT descriptors to better capture rotation invariance and spatial relationships, improving performance on MNIST datasets.

Contribution

It proposes a novel CNN architecture integrating SIFT descriptors instead of max-pooling to enhance rotation invariance and spatial feature capturing.

Findings

Improved accuracy on MNIST and FashionMNIST datasets

Enhanced rotation invariance in feature extraction

Combines CNN and SIFT advantages

Abstract

This paper presents a novel approach to exploit the distinctive invariant features in convolutional neural network. The proposed CNN model uses Scale Invariant Feature Transform (SIFT) descriptor instead of the max-pooling layer. Max-pooling layer discards the pose, i.e., translational and rotational relationship between the low-level features, and hence unable to capture the spatial hierarchies between low and high level features. The SIFT descriptor layer captures the orientation and the spatial relationship of the features extracted by convolutional layer. The proposed SIFT Descriptor CNN therefore combines the feature extraction capabilities of CNN model and rotation invariance of SIFT descriptor. Experimental results on the MNIST and fashionMNIST datasets indicates reasonable improvements over conventional methods available in literature.