Abstract
Filtering an image by eliminating noise and irrelevant details while preserving prominent structure edges is an important pre-processing task in many fields, such as image processing and computer vision. In this study, a novel approach called anisotropic joint trilateral rolling filter (AJTRF) is proposed for the smoothing of an image while preserving important structure edges. AJTRF works in an iterative manner and extends the joint bilateral filter to a joint trilateral filter by employing additional weights in eigen space. During filtering, the range weights are recalculated in each iteration, while the spatial Gaussian weights are computed in anisotropic orientations instead of isotropic orientations. Furthermore, the inner products between anisotropic orientations (expressed as eigenvectors) are used as weights to measure orientation similarities of pixels. Structure tensors at each pixel are calculated to determine the structure orientations, which are used as the anisotropic orientations. A range of experiments are conducted. The results demonstrate that the proposed AJTRF has stronger smoothing and structure preserving ability than established methods.
Highlights
Natural imagery contains many rich visual structures at different scales
Eliminating small scale textures while preserving the main structures can be achieved by an image smoothing filter
There exists a lot of traditional smoothing methods such as the box filter, weighted average filter, Gaussian filter, and bilateral filter [3] working in the spatial domain
Summary
Natural imagery contains many rich visual structures at different scales. in the majority of applications (including image abstraction, image segmentation, and object detection etc.) small-scale textures are less important while prominent structures are essential. There is the low-pass filter in the frequency domain that is used to smooth the image by removing high frequency components These traditional smoothing filters remove finer textures and noises, most of them (except for the bilateral filter) blur edges in the images, that is, the salient structures are lost. One of the purposes of image smoothing is to remove the unwanted small features that may represent finer textures, noises and small objects Another goal is to preserve the prominent structures such as region boundaries and large objects [2]. A new edge-preserving image smoothing filter called anisotropic joint trilateral rolling filter (AJTRF) is proposed. Unlike the original rolling guidance filter, which uses a Gaussian filter to remove small structures, the AJTRF combines the anisotropic Gaussian weight and inner product of eigenvectors of the structure tensor to remove small textures and restore the prominent structures.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
