Abstract:We propose a random convolutional neural network to generate a feature space in which we study image classification and retrieval performance. Put briefly we apply random convolutional blocks followed by global average pooling to generate a new feature, and we repeat this k times to produce a k-dimensional feature space. This can be interpreted as partitioning the space of image patches with random hyperplanes which we formalize as a random depthwise convolutional neural network. In the network's final layer we perform image classification and retrieval with the linear support vector machine and k-nearest neighbor classifiers and study other empirical properties. We show that the ratio of image pixel distribution similarity across classes to within classes is higher in our network's final layer compared to the input space. When we apply the linear support vector machine for image classification we see that the accuracy is higher than if we were to train just the final layer of VGG16, ResNet18, and DenseNet40 with random weights. In the same setting we compare it to an unsupervised feature learning method and find our accuracy to be comparable on CIFAR10 but higher on CIFAR100 and STL10. We see that the accuracy is not far behind that of trained networks, particularly in the top-k setting. For example the top-2 accuracy of our network is near 90% on both CIFAR10 and a 10-class mini ImageNet, and 85% on STL10. We find that k-nearest neighbor gives a comparable precision on the Corel Princeton Image Similarity Benchmark than if we were to use the final layer of trained networks. As with other networks we find that our network fails to a black box attack even though we lack a gradient and use the sign activation. We highlight sensitivity of our network to background as a potential pitfall and an advantage. Overall our work pushes the boundary of what can be achieved with random weights.

Randomized Distribution Feature for Image Classification.

Framework of Randomized Distribution Features for Visual Representation and Categorization.

Image Retrieval and Recognition Based on Generalized Local Distance Functions

Image feature detection as robust model fitting

Affine-Transformation-Invariant Image Classification by Differentiable Arithmetic Distribution Module

Metric Distribution to Vector: Constructing Data Representation via Broad-Scale Discrepancies

Linear Distance Coding for Image Classification

What Are the Distance Metrics for Local Features?

Binary Feature from Intensity Quantization and Weakly Spatial Contextual Coding for Image Search.

Simple Techniques Make Sense: Feature Pooling and Normalization for Image Classification

High-Order Distance-Based Multiview Stochastic Learning in Image Classification

On Data-Dependent Random Features for Improved Generalization in Supervised Learning

Image classification and retrieval with random depthwise signed convolutional neural networks

Metric forests based on Gaussian mixture model for visual image classification

Optimal Kernel Quantile Learning with Random Features

Modeling Outer Products of Features for Image Classification

A Novel Binary Feature from Intensity Difference Quantization Between Random Sample of Points

Generalized Learning Vector Quantization for Classification in Randomized Neural Networks and Hyperdimensional Computing

Convergence Theory of Generalized Distributed Subgradient Method with Random Quantization

Distributionally Robust Multiclass Classification and Applications in Deep Image Classifiers

Spatial Locality-Preserving Feature Coding for Image Classification