Description:FIELD OF THE INVENTION
This invention relates to Advanced Deep Learning Framework for CT Image Denoising Using Transform-Based Methods and Diffusion Modeling
BACKGROUND OF THE INVENTION
CT (Computed Tomography) imaging remains essential to medical diagnostics because it offers detailed cross-sectional pictures of the human body. The quality of these images becomes severely deteriorated by noise through low-dose CT (LDCT) scans that function to decrease patient radiation exposure during procedures. The presence of noise creates obstacles which conceal vital structures thus making diagnosis by clinicians much more difficult. The standard noise reduction techniques including filtering along with wavelet transforms fail to protect fine details and edges when removing noise since it results in vital diagnostic information being lost.
CT image denoising has become more effective through deep learning techniques which appear during recent years. Present deep learning-based approaches currently fail to integrate transform-based techniques and advanced diffusion process modeling because these techniques would help enhance denoising by better detecting image features and noise patterns. Optimizing the denoising process represents a difficult issue because achieving the correct balance between noise reduction and maintaining details remains delicate.
The study presents a modern deep learning structure uniting transform-based and diffusion modeling techniques for improving CT image denoising quality. These combined methods work to solve present imaging difficulties by improving noise elimination without damaging useful details to achieve better diagnostic outcomes in medical imaging.
CT image denoising products are commercially available through two solutions from NVIDIA Clara and Siemens Healthineers' Syngo Imaging software that employ deep learning algorithms to minimize noise levels. The image enhancement software OsiriX along with 3D Slicer utilizes wavelet transforms and filter-based denoising actions. CT scan noise reduction practices in commercial use implement deep learning models (such as U-Net and DnCNN) along with statistical methods that exist within medical imaging software to preserve vital details during denoising operations. Diffusion modeling together with transform-based techniques appears to be used only infrequently.
Various existing methods find it difficult to manage appropriate noise reduction techniques
with fine detail preservation particularly in cases of low-dose CT images. Traditional models tend to produce excessive smoothing of images although contemporary methods prove insufficient in handling intricate patterns of noise. Commercial noise reduction tools lack the incorporation of advanced methods including transform-based techniques and diffusion modeling which prevents them from achieving their best diagnostic accuracy level.
SUMMARY OF THE INVENTION
This summary is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention.
This summary is neither intended to identify key or essential inventive concepts of the invention and nor is it intended for determining the scope of the invention.
To further clarify advantages and features of the present invention, a more particular description of the invention will be rendered by reference to specific embodiments thereof, which is illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. The invention will be described and explained with additional specificity and detail with the accompanying drawings.
Diffusion modeling serves as the central mechanism in the proposed method to improve CT image denoising outcomes. Through simulation the diffusion process generates gradual smoothing functions which minimize high-frequency noise yet enable protection of diagnostic mandatory structural details. Through the implementation of diffusion, the model performs noise reduction without affecting essential diagnostic elements in medical images.
BRIEF DESCRIPTION OF THE DRAWINGS
The illustrated embodiments of the subject matter will be understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of devices, systems, and methods that are consistent with the subject matter as claimed herein, wherein:
FIGURE 1: SYSTEM ARCHITECTURE
The figures depict embodiments of the present subject matter for the purposes of illustration only. A person skilled in the art will easily recognize from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.
DETAILED DESCRIPTION OF THE INVENTION
The detailed description of various exemplary embodiments of the disclosure is described herein with reference to the accompanying drawings. It should be noted that the embodiments are described herein in such details as to clearly communicate the disclosure. However, the amount of details provided herein is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the present disclosure as defined by the appended claims.
It is also to be understood that various arrangements may be devised that, although not explicitly described or shown herein, embody the principles of the present disclosure. Moreover, all statements herein reciting principles, aspects, and embodiments of the present disclosure, as well as specific examples, are intended to encompass equivalents thereof.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms “a",” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and/or “including,” when used herein, specify the presence of stated features, integers, steps, operations, elements and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof.
It should also be noted that in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may, in fact, be executed concurrently or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
In addition, the descriptions of "first", "second", “third”, and the like in the present invention are used for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Thus, features defining "first" and "second" may include at least one of the features, either explicitly or implicitly.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which example embodiments belong. It will be further understood that terms, e.g., those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Diffusion modeling serves as the central mechanism in the proposed method to improve CT image denoising outcomes. Through simulation the diffusion process generates gradual smoothing functions which minimize high-frequency noise yet enable protection of diagnostic mandatory structural details. Through the implementation of diffusion, the model performs noise reduction without affecting essential diagnostic elements in medical images.
CT images are decomposed into multiple frequency components and scales through Wavelet and Curvelet and Shearlet transform analysis. Transforms specialize in detecting specific types of image features during their decomposition process. The detection capabilities of wavelets extend to identifying both edges and low-frequency information yet curvelets together with shearlets maintain superiority in processing complicated geometric structures. The image transforms separate noise components from true image structures by analyzing different scales and directional orientations. The initial preparation enhances denoising outcomes because it separates noise from vital image components effectively.
GANs operate as two competing networks to transform noisy images into their noise-free versions. During training the generator network develops expertise in producing natural images void of noise and the discriminator network verifies the unsuspiciousness of these generated images to authentic clean images. GANs optimize denoising framework performance through their ability to produce natural CT images of high quality which results in better PSNR, SSIM and maintains image edges. The integration of transforms with diffusion methods and GANs provides a solution to address image denoising challenges and preserves important image characteristics.
NOVELTY:
The proposed method implements deep learning methods together with multi-scale transforms and diffusion modeling to perform CT image denoising with superior noise reduction of structural elements. Through the integration of CNNs and GANs with transforms the solution significantly enhances image quality which results in remarkable improvements of PSNR, SSIM, FSIM and edge preservation when compared to previous approaches.
ADVANTAGES OF THE INVENTION
Better performance for noise removal occurs when multi-scale transforms (Wavelet and Shearlet) work alongside CNNs and GANs to save image details.
The structural edges along with the original information remain intact during Diffusion modeling since standard denoising techniques tend to produce blurring effects.
When applied together CNNs with GANs and transforms create a more powerful solution for denoising compared to using individual techniques.
, Claims:1. A computer-implemented method for denoising computed tomography (CT) images, the method comprising:
a) preprocessing noisy input CT images through noise reduction, normalization, and data augmentation steps;
b) decomposing the preprocessed images into multiple frequency components using a shearlet transform;
c) applying diffusion modeling including noise estimation and a diffusion process to reduce high-frequency noise while preserving structural details; and
d) evaluating the denoised images using PSNR, SSIM, and EPI metrics.
2. The method as claimed in claim 1, wherein the diffusion modeling comprises a generative adversarial network (GAN) framework, wherein a generator transforms noisy images into noise-free images, and a discriminator assesses the realism of the denoised images against clean references.
3. The method as claimed in claim 1, wherein the transform-based decomposition is performed using wavelet, curvelet, or shearlet transforms to detect features across different scales and orientations for distinguishing noise from image structures.
4. The method as claimed in claim 1, wherein the preprocessing enhances denoising outcomes by effectively separating noise from diagnostically relevant structures before application of the diffusion model.
5. The method as claimed in claim 1, wherein the GAN architecture is trained to optimize denoising performance by maximizing PSNR and SSIM metrics and preserving edge details in the CT images.