Abstract:Recently, many deep networks are proposed to learn hierarchical image representation to replace traditional hand-designed features. To enhance the ability of the generative model to tackle discriminative computer vision tasks (e.g. image classification), we propose a hierarchical deconvolutional network with two biologically inspired properties incorporated, i.e., non-negative sparsity and selectivity. First, we propose a single layer deconvolutional model with a raw image as input, attempting to decompose the input as a weighted sum of feature maps convolving with filters. Here, the filters are the model parameters common to all the inputs, while the feature maps and the summing weights are specific to the input. The non-negative sparsity is formulated as the /i-norm regularizer on the feature map, which is used to generate feature representations for image classification. And the selectivity is forced on the filters to make different filters active different inputs, through requiring the sparsity on the summing weights specifically. The two properties are summarized into an overall cost function, which can be solved with an alternatively iterative algorithm. Then, we build multiple layer deconvolutional network by stacking the single models, where the next-layer inputs are the results of a 3D max-pooling operation on the inferred feature maps of the front layer, and train the network in a greedy layer wise scheme. Finally, we explore the feature maps of each layer to generate the image representations and input them to a SVM classifier for the classification task. Experiments on two image benchmark datasets of Caltech-101 and Caltech-256 demonstrate the encouraging performance of our model compared with other deep feature learning models as well as some hand-designed features.

Learning Adaptive Filter Banks for Hierarchical Image Representation

Regularized Hierarchical Feature Learning with Non-negative Sparsity and Selectivity for Image Classification

Hierarchical Sparse Coding from a Bayesian Perspective.

Multiscale Dictionary Learning for Hierarchical Sparse Representation.

Hierarchical Feature Concatenation-Based Kernel Sparse Representations for Image Categorization

A Hybrid Wavelet Convolution Network with Sparse-Coding for Image Super-Resolution

WaveLBP based hierarchical features for image classification

An adaptive sparse representation for remote sensing image based on combination of wavelet and adaptive directional filter

Hierarchical Sparse Coding Based on Spatial Pooling and Multi-Feature Fusion.

Hierarchical Sparse Method with Applications in Vision and Speech Recognition

Learning Filter Bank Sparsifying Transforms

Convolutional Adaptive Denoising Autoencoders for Hierarchical Feature Extraction

Learning Hierarchical Dynamics with Spatial Adjacency for Image Enhancement

Generative Hierarchical Learning of Sparse FRAME Models

A Low-Complexity Bayesian Approach to Large-Scale Sparse Image Reconstruction with Structured Constraints

Hierarchical Discriminative Feature Learning for Hyperspectral Image Classification

A sparse infrastructure of wavelet network for nonparametric regression

Image Hierarchical Representations Models Based on Latent Dirichlet Allocation.

Image Super-Resolution via Hierarchical and Collaborative Sparse Representation

Robust Visual Tracking Via Hierarchical Convolutional Features-Based Sparse Learning

Hierarchical Discriminative Sparse Coding Via Bidirectional Connections