Deep Mean Maps

TL;DR

Deep Mean Maps (DMMs) integrate distributional feature representations with deep CNNs, enabling improved image classification by leveraging distributional patterns of features in a non-parametric manner.

Contribution

This paper introduces DMMs, a novel framework that combines distributional representations with deep learning, and demonstrates their effectiveness in image classification tasks.

Findings

DMMs improve CNN performance on real-world datasets.

DMMs successfully analyze distributional patterns in image data.

Implementation of DMMs is straightforward within existing CNN architectures.

Abstract

The use of distributions and high-level features from deep architecture has become commonplace in modern computer vision. Both of these methodologies have separately achieved a great deal of success in many computer vision tasks. However, there has been little work attempting to leverage the power of these to methodologies jointly. To this end, this paper presents the Deep Mean Maps (DMMs) framework, a novel family of methods to non-parametrically represent distributions of features in convolutional neural network models. DMMs are able to both classify images using the distribution of top-level features, and to tune the top-level features for performing this task. We show how to implement DMMs using a special mean map layer composed of typical CNN operations, making both forward and backward propagation simple. We illustrate the efficacy of DMMs at analyzing distributional patterns in image data in a synthetic data experiment. We also show that we extending existing deep architectures with DMMs improves the performance of existing CNNs on several challenging real-world datasets.