Image inhomogeneous mapping method based on grid deformation optimization

Abstract

The invention discloses an image inhomogeneous mapping method based on grid deformation optimization and belongs to the field of image processing. According to the image inhomogeneous mapping method based on grid deformation optimization, a fast image scaling technology based on content sensing and suitable for different user terminals is provided, images with fixed sizes can be displayed on different types of terminal equipment in a self-adaptive mode, important contents in an original image can be kept as much as possible, key objects are kept distortionless, and the import structure, namely the spatial position relation, in the image is kept unchanged so that optimal image observing quality can be guaranteed. According to the image inhomogeneous mapping method based on grid deformation optimization, a content-based image scaling optimal model is established by defining a series of energy functions capable of effectively controlling grid deformation, deformation of obvious key areas is avoided during image scaling, and in this way, user experience of browsing images on mobile digital equipment is improved; the quality of a scaling result can be effectively improved, and tensile deformation of the edges of the key areas can be avoided to a certain degree.

Description

technical field [0001] The invention relates to a content-aware-based image scaling technology in the technical field of image processing, in particular to the research and realization of an image non-uniform mapping method based on grid deformation optimization. Background technique [0002] With the rapid development of Internet technology and the continuous improvement of the hardware level of digital devices, terminal devices suitable for different networks are constantly updated, such as TVs, laptops, PDAs, mobile phones, etc., and they have different types of display sizes. For example, the current mainstream The screen ratio of large-screen smart phones is 5:3 or 16:9, the screen ratio of handheld tablet computers is mostly 16:10, and the ipad consistently maintains a screen ratio of 4:3 and so on. However, in most cases digital images are produced in a certain size and aspect ratio, but when displaying them on different types of digital devices, it is usually necessa...

AI Technical Summary

Problems solved by technology

Traditional image scaling methods are mainly divided into three types, one of which is to obtain an image of the target size through interpolation or downsampling of the original image, which is the so-called uniform scaling Scaling technology. , it tends to cause stretching of the main content of the image

Another method is the simple cropping technology, which obtains the target size by cutting off the edge content of the image, which will lead to the loss of more content information

The third method is to select the minimum scaling ratio of the image width or height as the overall scaling ratio to scale the original image proportionally. This method is also called "letterbox" letterboxing technology. Black borders are generated on the left and right sides, which cannot make full use of the limited screen resources of mobile devices

Because this type of algorithm combines the advantages of multiple methods, but at the same time it also introduces the shortcomings of various methods, then how to optimally combine various methods to determine the best operation sequence and operation amount requires an effective quality evaluation method To calculate the similarity between the source image and the target image to determine the switching point between different scaling methods

However, the time complexity of most multi-operation methods is high, so how to speed up the processing speed of multi-operation methods is also a problem that needs serious consideration.

Images