A Data Retrieval Method Based on Adaptive Binary Quantized Hash Coding

Abstract

The present invention provides a data retrieval method based on adaptive binary quantized hash coding, comprising: S1, selecting cluster centers in the data space to form a cluster center set, and assigning a Hamming space to each cluster center The binary code in the binary code is obtained to obtain the corresponding binary code set; S2, according to the obtained binary code set, update the cluster center set and the cluster center to which the data in the data space belongs, until the data space is consistent with the sea Align the bright space, and map all the data in the data space to the binary code corresponding to the cluster center to which it belongs, so as to complete the hash coding. A data retrieval method based on adaptive binary quantization hash coding provided by the present invention adopts an incomplete coding scheme to adaptively determine the number of cluster centers and the corresponding binary coding, thereby reducing quantization loss and improving hash coding. retrieval performance.

Description

Technical field [0001] The present invention relates to the field of computer data retrieval, and more specifically, to a data retrieval method based on adaptive binary quantization hash coding. Background technique [0002] At present, with the advent of the big data era, massive amounts of unstructured data such as images, videos, etc. are being generated on the Internet at all times. How to establish efficient retrieval algorithms for these data has become an urgent problem to be solved. In practical applications, people often need to query the most similar objects of a given sample, and the data is usually expressed as a feature vector obtained by feature extraction. For example, the local features of an image use a scale-invariant feature transform (SIFT) algorithm. The feature vector is obtained, and the global feature uses the generalized search tree algorithm (Generalized SearchTrees, GIST) to obtain the feature vector, so the problem is transformed into the nearest neigh...

AI Technical Summary

Problems solved by technology

[0004] However, most of the current methods try to establish a complete match between the data space and the coding space, and the data distribution usually does not satisfy the regular hypercube structure in the coding space, so these methods are difficult to maintain the spatial relationship, so the retrieval effect is not good

Images