Image similarity calculation method based on color quantization

Abstract

The invention discloses an image similarity calculation method based on color quantization. The image similarity calculation method comprises the following steps: step 1), converting a value of an HSVcolor space; step 2), performing quantization dimension reduction on the converted HSV color space; 3) giving images Pd and Pc to be matched; (4) equally dividing the image Pd and the image Pc into nparts from top to bottom; (6) obtaining the main color of the image Pdk according to the color quantization component graph, and recording the main color as Wdk, wherein k = 1, 2, 3,..., n, Wck, k =1, 2, 3,..., n; and step 8), giving a threshold T, and calculating whether the color Wdk is equal to the color Wck or not, wherein k = 1, 2, 3,..., n. According to the invention, image similarity calculation is carried out by using a simple and effective method in the fields of computer vision and image over-processing; by means of color space quantification and dimensionality reduction strategies, image matching accuracy is improved during similarity calculation, time expenditure is greatly saved, the requirement for real-time performance is met in the aspect of speed, and the method can be applied to various fields needing real-time similarity calculation.

Description

technical field [0001] The invention relates to various fields such as computer vision and image processing, and is specifically an image similarity calculation method based on spatial color quantization. Background technique [0002] With the advancement of social science and technology, the rapid and vigorous development of computer vision technology and image processing, the image similarity matching technology, as an important part of various fields of image processing, has already become an indispensable part of computer vision and image processing. A technology, and it is applied to various existing or emerging fields. [0003] Album clustering, face recognition matching, pedestrian re-recognition and vehicle re-recognition, etc., these emerging computer vision fields that have been relatively hot in recent years are all indispensable for the calculation of image similarity. Especially in recent years, with the rise of deep learning, the similarity calculation of imag...

AI Technical Summary

Problems solved by technology

[0007] (1) The time overhead is too large, and it takes a lot of time to calculate the similarity of two images, and it is difficult to play a role in some occasions that require real-time calculation;

[0008] (2) The accuracy of similarity matching is not high. The current similarity calculation methods are basically used in various matching and re-identification fields, but the accuracy of similarity calculation methods due to differences between images is not high;

Images