Image retrieval method based on CNN and dimension reduction

TL;DR

This paper presents an image retrieval approach combining CNN for feature extraction, multilinear PCA for dimension reduction, and binary hashing for fast retrieval, demonstrating improved performance over PCA-based methods on e-commerce datasets.

Contribution

Introduces a novel image retrieval method integrating CNN, multilinear PCA, and binary hashing to enhance retrieval accuracy and speed.

Findings

Outperforms PCA-based retrieval methods on e-commerce datasets

Effective reduction of feature dimensions while maintaining retrieval quality

Binary hashing enables rapid image retrieval

Abstract

An image retrieval method based on convolution neural network and dimension reduction is proposed in this paper. Convolution neural network is used to extract high-level features of images, and to solve the problem that the extracted feature dimensions are too high and have strong correlation, multilinear principal component analysis is used to reduce the dimension of features. The features after dimension reduction are binary hash coded for fast image retrieval. Experiments show that the method proposed in this paper has better retrieval effect than the retrieval method based on principal component analysis on the e-commerce image datasets.