Abstract:In the field of computer vision, fine-grained image retrieval is an extremely challenging task due to the inherently subtle intra-class object variations. In addition, the high-dimensional real-valued features extracted from large-scale fine-grained image datasets slow the retrieval speed and increase the storage cost. To solve above issues, existing fine-grained image retrieval methods mainly focus on finding more discriminative local regions for generating discriminative and compact hash codes, which achieve limited fine-grained image retrieval performance due to the large quantization errors and the confounding granularities and context of discriminative parts, i.e., the correct recognition of fine-grained objects mainly attribute to the discriminative parts and their context. To learn robust causal features and reduce the quantization errors, we propose a deep progressive asymmetric quantization (DPAQ) method based on causal intervention to learn compact and robust descriptions for fine-grained image retrieval task. Specifically, we introduce a structural causal model to learn robust casual features via causal intervention for fine-grained visual recognition. Subsequently, we design a progressive asymmetric quantization layer in the feature embedding space, which can preserve the semantic information and reduce the quantization errors sufficiently. Finally, we incorporate both the fine-grained image classification and retrieval tasks into an end-to-end deep learning architecture for generating robust and compact descriptions. Experimental results on several fine-grained image retrieval datasets demonstrate that the proposed DPAQ method performs the best for fine-grained image retrieval task and surpasses the state-of-the art fine-grained hashing methods by a large margin.

Progressive Similarity Preservation Learning for Deep Scalable Product Quantization

Beyond Product Quantization: Deep Progressive Quantization for Image Retrieval.

Pse: Mixed Quantization Framework of Neural Networks for Efficient Deployment

Bilinear Optimized Product Quantization for Scalable Visual Content Analysis.

Deep Product Quantization Module for Efficient Image Retrieval

Differentiable Optimized Product Quantization and Beyond

Shared Predictive Cross-Modal Deep Quantization.

Orthonormal Product Quantization Network for Scalable Face Image Retrieval

Deep Visual-Semantic Quantization For Efficient Image Retrieval

Entropy-based Deep Product Quantization for Visual Search and Deep Feature Compression

Unleashing the Full Potential of Product Quantization for Large-Scale Image Retrieval.

Scalable Image Retrieval by Sparse Product Quantization

Entropy-Optimized Deep Weighted Product Quantization for Image Retrieval

Matching-oriented Embedding Quantization for Ad-hoc Retrieval.

Spherical Centralized Quantization for Fast Image Retrieval

Deep Progressive Asymmetric Quantization Based on Causal Intervention for Fine-Grained Image Retrieval

Online Optimized Product Quantization

Mixed-precision Deep Neural Network Quantization With Multiple Compression Rates

Matching-oriented Product Quantization for Ad-hoc Retrieval

Towards Codebook-Free Deep Probabilistic Quantization for Image Retrieval.

Stacked Product Quantization For Nearest Neighbor Search On Large Datasets