Single channel image deformation system and method using anisotropic filtering

Abstract

The present invention discloses a method for executing circular symmetrical anisotropic filtering in the extended ellipse or rectangular track in the single-channel digital picture deformation and a system. When the filtering is executed the ellipse is found and adjusted, and the ellipse is approximate to the non-homogeneous image scaling function at the mapping position of the output pixel in the input image space. The linear transformation from the ellipse to the unit circle in the input image space is confirmed to calculate the input image radius in the track and a corresponding wave filter coefficient as the function of the radius. Compromise is executed between the image quality and the processing speed to confirm the track shape. In one embodiment the profiles of the smooth and deformed parts are combined together to generate more clear or detail-reinforced output image. The method and system of the invention can generate lifelike output image without sawtooth distortion and at the same time the definition of the input image can be remained or reinforced.

Description

technical field [0001] The invention relates to a filtering method in digital image transformation, more specifically, to an anisotropic filtering in signal resampling of a single-channel digital image deformation system. Background technique [0002] A digital image is basically a two-dimensional array of digital data, where each element represents a pixel of the digitized image. Pixel data includes various luminance components such as red, green, and blue (RGB). The display system electronically assigns brightness values ​​to these various components based on the entered digital data. However, the displayed image may suffer from various optical and geometric distortions such as lens distortion such as pincushion and barrel distortion, projected keystone distortion, and the like. One method of compensating for these distortions is digital image warping, or "non-uniform scaling". Image warping can be defined as a process of dynamically resampling an evenly spaced input im...

AI Technical Summary

Problems solved by technology

Second, segmentation itself introduces geometric distortions

Third, dual-channel filtering is prone to jagging artifacts, especially when geometric transformations have rotational components

[0013] However, these and other existing systems have difficulty achieving a trade-off between aliasing and sharpness due to their impulse responses resulting in a single frequency response

Images