Optimized Feature Space Learning for Generating Efficient Binary Codes for Image Retrieval
This work addresses the challenge of efficient image retrieval for both single- and multi-labeled datasets, though it is incremental as it builds on existing techniques like LDA, CCA, and ITQ.
The paper tackles the problem of high-dimensional feature vectors in image retrieval by learning an optimized low-dimensional feature space that minimizes intra-class variance and maximizes inter-class variance, using a method that combines Linear Discriminant Analysis and Canonical Correlation Analysis for single- and multi-labeled images, and achieves competitive mean average precision results on state-of-the-art datasets.
In this paper we propose an approach for learning low dimensional optimized feature space with minimum intra-class variance and maximum inter-class variance. We address the problem of high-dimensionality of feature vectors extracted from neural networks by taking care of the global statistics of feature space. Classical approach of Linear Discriminant Analysis (LDA) is generally used for generating an optimized low dimensional feature space for single-labeled images. Since, image retrieval involves both multi-labeled and single-labeled images, we utilize the equivalence between LDA and Canonical Correlation Analysis (CCA) to generate an optimized feature space for single-labeled images and use CCA to generate an optimized feature space for multi-labeled images. Our approach correlates the projections of feature vectors with label vectors in our CCA based network architecture. The neural network minimize a loss function which maximizes the correlation coefficients. We binarize our generated feature vectors with the popular Iterative Quantization (ITQ) approach and also propose an ensemble network to generate binary codes of desired bit length for image retrieval. Our measurement of mean average precision shows competitive results on other state-of-the-art single-labeled and multi-labeled image retrieval datasets.