CNN-DBN: Quality Assessment and Optimization of Content-Based Image Retrieval Services

The quality assessment of image retrieval services can master the quality of retrieval results in real-time, and provide valuable information for users. Therefore, the quality assessment of image retrieval, especially content-based image retrieval(CBIR) services, has attracted the attention of researchers. However, the traditional quality assessment approach of CBIR services had some problems, for example, they ignored the high-level semantics information of the image, used shallow learning models, etc. To address these problems, this paper proposes a quality assessment approach of image retrieval, namely CNN-DBN, it extracts image features by Convolutional Neural Network(CNN) and establishes a quality assessment model using Deep Belief Network(DBN). Based on the quality assessment results, this paper proposes a scheme to optimize the quality of CBIR. To evaluate the performance of this approach, design a set of experiments to verify the performance and practicability of the proposed approach. To a certain extent, it can evaluate the quality of image retrieval and optimize the quality of retrieval results, demonstrate the importance of image retrieval quality assessment for users.