SSD Net: Toward deep network models based on dissimilarity metrics

摘要

Most artificial neural networks that originate from the perceptron use the inner product as the basic operation to calculatepattern similarities. Unlike them, we propose a novel hierarchical network model based on a pattern dissimilarityoperation using a popular dissimilarity metric: sum of squared differences. Our model is differentiable and end-to-endtrainable. We provide a description of the basic formulation and network architecture of the proposed method. Then weapply our method to image classification tasks using public datasets for performance comparison. Although our methoddoes not outperform the same size of convolutional neural network in terms of classification accuracy, it demonstratesthat comparable performance can be obtained.