Bilateral Filtering Algorithm Research Articles

A Fast Point-Sampled Model Denoising Algorithm for Product Design

Denoising is an essential step in creating perfect point-sampled models, which are widly used in the field of product Design. As the bilateral mesh deoising method by Fleishman has been extended to 3D point-sampled model, which is not efficient enough. For the reason, a method is proposed that using the quasi-Cauchy kernel to replace the Gauss kernel used in the bilateral filtering algorithm, so that the efficiency of the algorithm is improved. Experiments show that the method can eliminate the noise on the surface of the point-sampled model effectively and meanwhile the sharp features of the surface are preserved well．

Image Restoration Method Based on Pre-Filtering for Cone-Beam Computed Tomography

For the problem of image quality degradation of cone-beam Computed Tomography (CBCT) based on flat panel detector (FPD), a constrained least squares iteration (CLSI) restoration method based on pre-filtering is proposed. Firstly, the original projected images are denoised with bilateral filtering algorithm. Then, the denoised projected images are restored with CLSI. Finally, the final restored images are obtained by adding the noise images, which got by subtracting the projected images before and after denoising, to the restored images. The experimental results show that the method well inhibits the noise amplification phenomenon in image restoration, and increases the edge sharpness and contrast-to-noise ratio (CNR) of the projected images and slice images. The CBCT image quality is significantly improved with this method.

Bilateral Filtering for Image Processing Based on Pulse Coupled Neural Networks

To effectively remove Gaussian noise in images, a novel bilateral filtering algorithm based on improved pulse coupled neural networks (PCNN) is introduced. Firstly, smooth restraining factor and adaptive linking strength are adopted. They are combined with the synchronous fire of similar neuron in improved PCNN from the aspect of the characteristics of image Gaussian noise. Secondly, the noisy image is processed by using PCNN pre-de-noise iteration. The extreme value noise is removed and the time matrix that can reflect image spatial–temporal information is also formed. Finally, the time matrix is operated in bilateral filtering which is given adaptive improvement and de-noise application. The results of experiments show that the novel algorithm can effectively remove noise in smooth region in the case of good preserving image edges and details; the de-noised images have good subjective vision effects and objective quality. Meanwhile, this algorithm presents high signal-to-noise ratio, strong capability to reduce any and has good adaptability.

TU-A-201B-02: 20-Fold Dose Reduction Using a Gradient Adaptive Bilateral Filter: Demonstration Using in Vivo Animal Perfusion CT

Purpose: To demonstrate the dose saving potential of a fast and practical algorithm based on a novel Gradient Adaptive Bilateral filtration (GABI) algorithm, applied to in vivo extremely low dose animal perfusion CT scans. Methods: the GABI algorithm is mainly applied in the time dimension to avoid losing spatial resolution and noise power spectra characteristics (texture of CT images). It expands the bilateral filter algorithm by non-locally adapting its strength based on a temporal gradient applied to the image dataset, and hence preserving contrast fidelity over time. We tested the GABI on two renal perfusion animal experiments which used a reference dose of 160 mAs (full dose FD), and 16 mAs and 8 mAs (1/10th and 1/20th of the radiation dose). Results: we demonstrated that the GABI algorithm preserves the temporal fidelity, as measured by the time attenuation curves (TACs), while dramatically reducing image noise from very low dose acquisitions, keeping comparable spatial resolution, and preserving CT noise texture. The background SNR for the FD pig A, FD pig B, 1/10th dose pig A, 1/20th dose pig B, 1/10th pig A + GABI, 1/20th pig B + GABI were 4.5, 4.4, 1.4, 0.85, 3.8 and 2.9 respectively. Higher background SNR improvements are possible, however we were constrained to keep very high fidelity to the TACs and hence perfusion parameters. Conclusions: the GABI algorithm provides a practical framework to highly reduce the dose in CT time series studies. The method is fast since it is image-space based and non-iterative. We demonstrated in real animal data, that in high SNR tasks as renal perfusion the gain in dose reduction can be as high as 20-fold with no practical detriment of perfusion information and image quality comparable to a full dose image acquisition. Further investigation for parameter optimization is guaranteed.

Simplification of Point Set Surfaces using Bilateral Filter and Multi-Sized Splats

We have developed a new algorithm that addresses the two major issues that are critical for the use of point-based rendering in real-world applications: rendering performance and rendering quality. The proposed algorithm improves rendering performance by reducing the number of points to be rendered. We generated points with different sizes in order to improve the quality of the point-splatting-type rendering. In this paper, we propose a feature-preserved simplification algorithm for point-sampled surfaces that generates models with different levels of details. The proposed method automatically balances the sampling density and point sizes. Our algorithm iteratively reduces the number of points using a bilateral filtering algorithm. We validate our method on the basis of the rendering results of different models with different resolutions.

Multitone block truncation coding

Block truncation coding (BTC) is an efficient image coding method with wide applications. However, most BTC methods produce only binary image blocks for display. Thus the limited representation power of binary images has been a major hindering factor for BTC performance improvement. A new multitone block truncation coding (MBTC) approach is presented, which breaks this limit by producing multitone pixel values. This approach uses a modified bilateral filtering algorithm to effectively extract the spatial information inherent in the binary coded image blocks. The proposed method significantly improves the code rate and the reconstructed image quality while keeping the processing efficiency among the best published results, making it suitable for high speed real time image and video communication and display.

A Fast Approximation to the Bilateral Filter for ITK

This document describes an implementation of an alternative bilateral filtering algorithm in ITK. This class implements the fast bilateral filter algorithm in ITK by taking the input image and organizing it into the required data structure as an ITK Image with dimension one greater than the original image. The DiscreteGaussianImageFilter is then used to blur the higher dimensional image. The discrete Gaussian image filter was chosen because the kernel width used should usually be small due to the down-sampling involved when using this technique. As described in the ITK documentation, this class should outperform the recursive Gaussian algorithm due to the small kernel size DiscreteGaussianImageFilter.After the Gaussian blur has been completed the data will be interpolated and written to the output image.By incorporating this class into the ITK framework, this more efficient algorithm will be made available to all ITK users.

On the Convergence of Bilateral Filter for Edge-Preserving Image Smoothing

The bilateral filter represents a wide group of nonlinear filters for edge-preserving image smoothing. In this work, we study the convergence properties of the bilateral filter algorithm. The understanding is established that the bilateral filter is an optimization procedure. We demonstrate that the bilateral filter is equivalent to minimizing a robust cost criterion using iterative reweighting, which is a good approximation to the very fast but unstable Newton's method. Further, the results of the analysis allow us to derive an improved hybrid smoothing scheme with concerns of computational efficiency and edge preservation.