A minimal model for classification of rotated objects with prediction of the angle of rotation

Abstract

In classification tasks, the robustness against various image transformations remains a crucial property of the Convolutional Neural Networks (CNNs). It can be acquired using the data augmentation. It comes, however, at the price of the risk of overfitting and a considerable increase in training time. Consequently, other ways to endow CNN with invariance to various transformations-and mainly to the rotations-is an intensive field of study. This paper presents a new reduced rotation invariant classification model composed of two parts: a feature representation mapping and a classifier. We provide an insight into the principle and we prove that the proposed model is trainable. This model is smaller in terms of trainable parameters than similar approaches, and has angular prediction capabilities. We illustrate the results on the MNIST and CIFAR-10 datasets. On MNIST, we i) achieve the state of the art of classification on MNIST-rot (with training on MNIST-rot), and ii) improve the results of classification on MNIST-rot (with training on upright MNIST). When trained on CIFAR-10 with upright samples and tested with rotated samples we improve by 20% the state of the art classification results. In all cases, we can predict the rotation angle.

GenHarris-ResNet: A Rotation Invariant Neural Network Based on Elementary Symmetric Polynomials

“…Comparison of the test error obtained on MNIST-Rot when training on MNIST with the five best results from the paper[23]. Networks: RP_RF_1_32[24], Spherical CNN[25], ORN-8 (ORAlign)[26], RED-NN[23], Covariant CNN[27].…”