Abstract:The state-of-the-art image classification models, generally including feature coding and pooling, have been widely adopted to generate discriminative and robust image representations. However, the coding schemes available in these models only preserve salient features which results in information loss in the process of generating final image representations. To address this issue, we propose a novel spatial locality-preserving feature coding strategy which selects representative codebook atoms based on their density distribution to retain the structure of features more completely and make representations more descriptive. In the codebook learning stage, we propose an effective approximated K-means with cluster closures to initialize the codebook and independently adjust the center of each cluster of the dense regions. Afterwards, in the coding stage, we first define the concept of “density” to describe the spatial relationship among the code atoms and the features. Then, the responses of local features are adaptively encoded. Finally, in the pooling stage, a locality-preserving pooling strategy is utilized to aggregate the encoded response vectors into a statistical vector for representing the whole image or all the regions of interest. We carry out image classification experiments on three commonly used benchmark datasets including 15-Scene, Caltech-101, and Caltech-256. The experimental results demonstrate that, comparing with the state-of-the-art Bag-of-Words (BoW) based methods, our approach achieves the best classification accuracy on these benchmarked datasets.

Modeling Outer Products of Features for Image Classification

Image Classification Method by Combining Multi-features and Sparse Coding

Image Feature Extraction Method Based On Sparse Coding Framework Of Hypergraph

Effects on Statistical Features for Image by Quantization

Scalable Image Retrieval by Sparse Product Quantization

Deep Product Quantization Module for Efficient Image Retrieval

Nonnegative sparse locally linear coding

Learning Nonlinear Statistical Regularities in Natural Images by Modeling the Outer Product of Image Intensities.

An Efficient Hybrid Feature For Edge-Preserving Based On Block Truncation Coding And Tree-Structured Vector Quantization With Edge Orientation Classification Of Bit-Planes

Sparse coding for image feature extraction and image recognition

Kernelized Product Quantization.

Local Hybrid Coding for Image Classification

Linear Distance Coding for Image Classification

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

Beyond Product Quantization: Deep Progressive Quantization for Image Retrieval.

Orthonormal Product Quantization Network for Scalable Face Image Retrieval

Scalar Quantization as Sparse Least Square Optimization

VQ Image Coding Using Sub-Vector Techniques.

Spatial Locality-Preserving Feature Coding for Image Classification

Image Quality Assessment Method Based on Nonlinear Feature Extraction in Kernel Space

Progressive Similarity Preservation Learning for Deep Scalable Product Quantization