Browse Prior Art Database

DENOISING TECHNIQUE FOR COMPUTED TOMOGRAPHY (CT) IMAGES

IP.com Disclosure Number: IPCOM000245118D
Publication Date: 2016-Feb-10
Document File: 5 page(s) / 403K

Publishing Venue

The IP.com Prior Art Database

Abstract

A technique for denoising computed tomography (CT) images is disclosed. The technique described herein is a single pass non-linear hybrid bilateral non-local mean denoising technique that non-local mean filter (NLM) with bilateral filter (BLF) to achieve improved denoising capabilities without losing structural information from the CT image.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

DENOISING TECHNIQUE FOR COMPUTED TOMOGRAPHY (CT) IMAGES

BACKGROUND

 

The present disclosure relates generally to computed tomography (CT) imaging and more particularly to a denoising technique for CT images.

Medical images are generally noisy due to physical components of image acquisition process. In CT imaging there is a scope to adapt image quality and dose. Reduction in radiation dose affects the quality of image and is generally responsible for image noise. There are various techniques conventionally used for denoising. Two conventional denoising techniques used frequently are bilateral filtering and non-local mean filter.  

Figure 1 depicts technique and formula for bilateral filter (BLF). The BLF is a non-linear, edge-preserving and noise-reducing smoothing filter for images. Intensity value at each pixel in an image is replaced by a weighted average of intensity values from nearby pixels.

Figure 1

Figure 2 depicts formula for non-local mean filter (NLM). According to this formula the denoised value at i is the mean of all the values at all the points whose Gaussian neighborhood is as the neighborhood of i.

Figure 2

However, efficient image denoising of images remains a challenge. In spite of sophistication of the conventional techniques, most algorithms do not provide the desired level of results. Most techniques show good performance when the image model corresponds to algorithm assumptions of the technique, but fail in general and create artifacts or remove fine structures from the image.

It would be desirable to have an improved denoising technique for computed tomography images.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 depicts technique and algorithm for edge preserving bilateral filter (BLF).

Figure 2 depicts algorithm for non-local mean filter (NLM).

Figure 3 is a graphical representation of a denoising technique described herein.

Figure 4 depicts formulas used for denoising according to an embodiment of the technique.

Figure 5 is parameters table for comparison images of heart processed using conventional and the proposed denoising technique.

Figure 6 depicts an example of a comparison between 4 images of heart, a raw image with heavy noise, an image processed using the NLM filter, an image processed us...