Local descriptor compression method and device

Abstract

A method for compressing a local feature descriptor is provided in an embodiment of the present invention. The method comprises: selecting from a target image one local feature descriptor or a plurality thereof; according to a preset codebook, carrying out multistage vector quantization on the selected local feature descriptor, and quantizing same as a feature codestream, said codestream comprising serial numbers of codewords obtained by means of the multistage vector quantization. A device for compressing a local feature descriptor and a storage medium are also provided in embodiments of the present invention.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to a local descriptor compression method and device. Background technique [0002] With the popularization of smart devices and mobile Internet, the application of mobile visual search is becoming more and more extensive. Among them, the relevant techniques of local descriptors have been widely used in mobile visual search. In the existing technology, image retrieval often needs to query and match through local descriptors, but due to limitations such as bandwidth, memory, and computing power, local descriptors often need to be compressed to achieve the purpose of compact expression; at the same time , the scale of the image database is getting larger and larger, and it is also necessary to compress the local descriptors of the database image to reduce disk consumption. [0003] Currently, image retrieval methods based on mobile devices include the following two: [0004]...

AI Technical Summary

Problems solved by technology

[0006] However, the above-mentioned first image retrieval method has complex calculations and huge codebooks in the local descriptor compression process, making the entire retrieval process take a long time and occupy a large memory space. At the same time, the compressed local descriptors do not have the compression ratio and The controllability of compression accuracy often leads to loss of information or excessive bandwidth usage, resulting in greater compression loss, resulting in poor retrieval results or long retrieval response time

Therefore, the image compression algorithm has limited capabilities and the image retrieval method has a high amount of calculation. For low-performance mobile devices, the process of extracting local descriptors will be very time-consuming, which will seriously affect the response time of the server and reduce the retrieval efficiency.

[0007] The second above-mentioned image retrieval method, due to the limited compression capability of the existing image compression method, the compressed image for transmission loses information, or affects the performance and transmission time of image retrieval due to taking up too much bandwidth

[0008] Therefore, in the prior art image retrieval methods, in the case of limited computing power, memory, and mobile bandwidth of mobile devices, the speed and accuracy of image retrieval will be greatly limited, resulting in a decrease in user experience

Images