A Method of Image Object Relocation Based on Local Uniform Scaling

Abstract

The invention discloses an image target relocation method based on local uniform scaling. First, the source image I is decomposed into a cartoon part u and a texture part v; then the high-order statistic HOS of each pixel on the cartoon part u is calculated to detect significant The edge is used to obtain the energy map of the source image, and according to the position of the salient target on the energy map, the source image is divided into three blocks: I1, I2 and I3, where I1 and I3 are the background parts and I2 is the salient area; Finally, the background parts I1, I3 and the salient area I2 are scaled according to different scaling ratios to finally obtain the target image. The present invention can accurately detect salient targets and large-scale structures of images, and the scaling speed is faster than other content-based scaling algorithms. It can well protect the large-scale structure of the image from being destroyed and maintain the large-scale structure. Smoothness and continuity of structure.

Description

technical field [0001] The invention relates to an image processing technology, in particular to an image object relocation method. Background technique [0002] In recent years, digital images have become more and more widely distributed in all aspects of the media field, and the development of mobile multimedia such as mobile phones and tablet computers has enriched the forms we browse images and videos. In order for images to be displayed normally on these devices, it is usually necessary to change the size of the image to fit the screen resolution of these devices. This process of changing the image size or resolution for a certain need is usually called image target relocation or image zoom. At present, the most commonly used image scaling method is uniform scaling, that is, through downsampling or image interpolation to achieve the purpose of reducing or enlarging the image, but because it fails to take into account the content of the image while changing the image si...

AI Technical Summary

Problems solved by technology

[0006] 3. The target image should have no obvious visual distortion or distortion

As the earliest and most common image scaling method, uniform scaling has many advantages such as fast speed and high matching rate of the source image. Its biggest defect is that it cannot take care of the image content and cause image distortion during the scaling process.

The subsequently proposed and widely studied seam carving algorithm is sensitive to salient objects, but it is easy to cause damage to large-scale structures due to excessive seam removal.

Subsequent researchers made some improvements to the seam carving algorithm, and proposed a forward energy algorithm, which minimizes the newly introduced energy after the image is removed from the seam, making the deformation caused by image processing relatively imperceptible.

Images