Abstract:Image representation based on sparse coding generalizes the bag of words model. Although it reduces the reconstruction error for local features to achieve the state-of-the-art image classification performance, the large computational cost hinders the application of sparse coding-based image features. In this paper, we propose approximating a sparse code using the output of a simple neural network. The resulting parameter learning model for the neural network automatically incorporates non-negative and shift-invariant constraints, leading to an efficient normalized non-negative sparse coding (N 3 SC) sparse encoder. Without the use of the traditional iterative process to solve the sparse coding objective, the sparse encoder directly “converts” each local feature into a sparse code. We also introduce a method for training the encoder based on the auto-encoder method. In addition, we formally propose the corresponding sparse coding scheme called N 3 SC, which enforces both the non-negative constraint and the shift-invariant constraint in addition to the traditional sparse coding criteria. As demonstrated by several experiments, the obtained N 3 SC encoder requires only 3%–10% of the processing time for image feature extraction compared with the standard sparse coding scheme. At the same time, the features extracted using the exact solutions of the N 3 SC coding scheme and the N 3 SC encoder offer superior image classification accuracy compared to the accuracy of many existing sparse coding-based representations.

Sparse Autoencoder for Social Image Understanding.

Laplacian Sparse Dictionary Learning for Image Classification Based on Sparse Representation

Deep Marginalized Sparse Denoising Auto-Encoder for Image Denoising

Understanding the Aesthetic Styles of Social Images.

The Face Recognition Based on the Sparse Autoencoder

Locality-Constrained Sparse Auto-Encoder for Image Classification

Image Super-Resolution Algorithm Based on an Improved Sparse Autoencoder

Zero Shot Learning Via Low-rank Embedded Semantic AutoEncoder

A New Deep Sparse Autoencoder for Community Detection in Complex Networks

Stacked Convolutional Sparse Auto-Encoders for Representation Learning

Learning Robust Uniform Features for Cross-Media Social Data by Using Cross Autoencoders.

Facial Expression Recognition Via Learning Deep Sparse Autoencoders.

Normalized Non-Negative Sparse Encoder for Fast Image Representation

Image Object Recognition via Deep Feature-Based Adaptive Joint Sparse Representation

Facial expression recognition and generation using sparse autoencoder

A Deep Learning Method Combined Sparse Autoencoder with SVM

Sparse Clustering Algorithm Based on Multi-Domain Dimensionality Reduction Autoencoder

Learning Deep Representations Using Convolutional Auto-Encoders with Symmetric Skip Connections

Fuzzy Sparse Autoencoder Framework for Single Image Per Person Face Recognition.

A Sparse Deep Learning Model for Privacy Attack on Remote Sensing Images

Sparse Autoencoder Based Spatial Pyramid Facial Feature Learning