Abstract:Learning effective feature representations and similarity measures are crucial to the retrieval performance of a content-based image retrieval (CBIR) system. Despite extensive research efforts for decades, it remains one of the most challenging open problems that considerably hinders the successes of real-world CBIR systems. The key challenge has been attributed to the well-known ``semantic gap'' issue that exists between low-level image pixels captured by machines and high-level semantic concepts perceived by human. Among various techniques, machine learning has been actively investigated as a possible direction to bridge the semantic gap in the long term. Inspired by recent successes of deep learning techniques for computer vision and other applications, in this paper, we attempt to address an open problem: if deep learning is a hope for bridging the semantic gap in CBIR and how much improvements in CBIR tasks can be achieved by exploring the state-of-the-art deep learning techniques for learning feature representations and similarity measures. Specifically, we investigate a framework of deep learning with application to CBIR tasks with an extensive set of empirical studies by examining a state-of-the-art deep learning method (Convolutional Neural Networks) for CBIR tasks under varied settings. From our empirical studies, we find some encouraging results and summarize some important insights for future research.

Region-Based Semantic Similarity Propagation for Image Retrieval.

Semantic Image Retrieval Based on Multiple-Instance Learning

Research on high-level semantic image retrieval

Multi-model Similarity Propagation and Its Application for Web Image Retrieval.

Query aware visual similarity propagation for image search reranking.

A Novel Semantics-Based Image Retrieval Method Using Similarity Measure of Multi-Level Semantics

Mixed Propagation for Image Retrieval

A Region-based Image Similarity Computation Approach

Collaborative Learning Between Visual Content and Hidden Semantic for Image Retrieval

Learning a hybrid similarity measure for image retrieval.

Image Retrieval Based on Fuzzy Semantic Relevance Matrix

Image Semantic Annotating Method Based on Region Features Relevancy

Image Interactive Retrieval Model Based on Integrated Region Similarity

Collaborative Similarity Metric Learning for Semantic Image Annotation and Retrieval.

Deep Learning for Content-Based Image Retrieval: A Comprehensive Study

Remotely Sensed Image Retrieval Based on Region-Level Semantic Mining

Region-based Image Retrieval Using Semantic Mining

Learning similarity for image retrieval with locally spatial information feedback

Multilabel Neighborhood Propagation for Region-Based Image Retrieval

A Method Based on Semantic Network for Image Retrieval

Semantic Sensitive Region Retrieval Using Keyword-Integrated Bayesian Reasoning