WO2012037091A1

WO2012037091A1 - Feature preservation in colon unfolding

Info

Publication number: WO2012037091A1
Application number: PCT/US2011/051351
Authority: WO
Inventors: Wei Hong
Original assignee: Siemens Corp
Current assignee: Siemens Corp
Priority date: 2010-09-17
Filing date: 2011-09-13
Publication date: 2012-03-22
Anticipated expiration: 2013-03-17
Also published as: WO2012037091A4

Abstract

A volumetric data unfolding method includes applying a histogram based intensity classification to a volumetric data including an object of interest for identifying the object of interest and a material in the object of interest (101), segmenting the object of interest from the volumetric data (102), determining a centerline of the object of interest (103), casting a plurality of rays from the centerline to determine a surface of the object of interest (104), wherein the plurality of rays ignore the material in the object of interest, resampling the centerline at a plurality of sampling points along the plurality of rays, fitting the surface of the object of interest with B-splines using the sampling points to determine a B-spline surface (105), wherein topological noise is filtered, determining a feature of the B-spline surface to be preserved (106), and unfolding surface of the object of interest (107) to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B-spline surface and mapping the surface of the B-spline surface to a planar surface, wherein an unfolded surface of the object of interest is displayed as a two-dimensional image.

Description

FEATURE PRESERVATION IN COLON UNFOLDING

CROSS-REFERENCE TO RELATED APPLICATION This is a non-provisional application claiming the benefit of U.S. provisional application serial number 61/383,893, filed September 17, 2010, the contents of which are incorporated by reference herein in their entirety.

BACKGROUND

1. Technical Field

The present disclosure relates to image editing, and more particularly to a volumetric data unfolding method.

2. Discussion of Related Art

Colorectal cancer is a leading cause of cancer related deaths. Colorectal cancer accounts for approximately 945,000 new cases and 500,000 deaths worldwide each year. Most colorectal cancers begin as a polyp, which is a small, harmless growth in the wall of the colon. As a polyp gets larger, it can develop into a cancer that grows and spreads. Early detection of colon cancer is the key ot a good prognosis. It can take from 10 to 15 years for an adenomatous polyp to become an invasive cancer. Thus, there is a considerable time for detection and clinical intervention if the proper screening methods are used.

To encourage people to participate in screening programs, virtual colonoscopy (VC), also known as computed tomographic colonography (CTC), has been proposed and developed to detect colorectal polyps using CT images of a patient's abdomen and a virtual fly-through visualization system that allows physicians to navigate within a three dimensional (3D) model of the colon searching for polyps. VC has shown promising results for colorectal cancer screening. However, because of the length of the colon, inspecting the entire colon wall is time consuming and prone to errors. Moreover, polyps behind folds may be hidden, which results in incomplete examinations. Virtual colon unfolding is an efficient visualization technique for polyp detection, in which the entire inner surface of the colon is displayed as a single 2D image. However, due to the complexity of the colon surface, the colon surfaces reconstructed from a CT data set usually have complicated topologies caused by the noise and inaccuracy of the reconstruction methods. This topological noise makes processing algorithm complicated and unstable. Moreover, remnants of stool and residual floods within the colon lumen make the situation even worse. Another challenge is that when 3D surfaces are mapped to two-dimensional (2D) planar surfaces, either area distortion or angle distortion will be introduced.

BRIEF SUMMARY

According to an embodiment of the present disclosure, a volumetric data unfolding method includes applying a histogram based intensity classification to a volumetric data including an object of interest for identifying the object of interest and a material in the object of interest, segmenting the object of interest from the volumetric data, determining a centerline of the object of interest, casting a plurality of rays from the centerline to determine a surface of the object of interest, wherein the plurality of rays ignore the material in the object of interest, resampling the centerline at a plurality of sampling points along the plurality of rays, fitting the surface of the object of interest with B-splines using the sampling points to determine a B-spline surface, wherein topological noise is filtered, determining a feature of the B-spline surface to be preserved, and unfolding surface of the object of interest to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B- spline surface and mapping the surface of the B-spline surface to a planar surface, wherein an unfolded surface of the object of interest is displayed as a two-dimensional image.

According to an embodiment of the present disclosure, a volumetric data unfolding method includes applying a histogram based intensity classification to a volumetric data including an object of interest for identifying the object of interest and a material in the object of interest, segmenting the object of interest from the volumetric data, determining a centerline of the object of interest, casting a plurality of rays from the centerline to determine a surface of the object of interest, wherein the plurality of rays ignore the material in the object of interest, resampling the centerline at a plurality of sampling points along the plurality of rays, determining a feature of the surface to be preserved, and unfolding surface of the object of interest to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B-spline surface and mapping the surface of the B-spline surface to a planar surface.

According to an embodiment of the present disclosure, a volumetric data unfolding system includes a processor configured to cast a plurality of rays from a centerline of an object of interest to determine a surface of the object of interest, wherein the plurality of rays ignore at least one type of tagged material in the object of interest, the processor further configured to resample the centerline at a plurality of sampling points along the plurality of rays, the processor further configured fit the surface of the object of interest with B-splines using the sampling points to determine a B-spline surface, wherein topological noise is filtered, the processor further configured determine a feature of the B-spline surface to be preserved, and the processor further configured unfold surface of the object of interest to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B-spline surface and mapping the surface of the B-spline surface to a planar surface, and a memory configured to store the unfolded surface.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the present disclosure will be described below in more detail, with reference to the accompanying drawings:

FIG. 1 A is a flow chart showing a method for unfolding image data according to an embodiment of the present disclosure;

FIG. IB is a flow chart showing a method for unfolding image data according to an embodiment of the present disclosure; FIG. 1C is a flow chart showing a method for unfolding image data according to an embodiment of the present disclosure; and

FIG. 2 is a diagram of a system for unfolding image data according to an embodiment of the present disclosure.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

According to an embodiment of the present disclosure, an inner surface of a colon lumen may be virtually unfolded from CT data, even when fecal tagging CT data is used.

According to an embodiment of the present disclosure, an exemplary volumetric data unfolding method does not require the input data to be clean. It uses a ray casting method to find the inner colon surface even when there are remnants of stool and residual floods inside the colon lumen.

According to an embodiment of the present disclosure, an exemplary volumetric data unfolding method may generate a topology simple colon surface for stable colon unfolding. In the application, a topology simple surface is free of topology noise, such as that typically generated by a marching cube method. The topology noise is expressed as tunnels or handles in the 3D triangle mesh.

According to an embodiment of the present disclosure, an exemplary volumetric data unfolding method can reduce distortion and preserve detected features, such as potential polyps.

According to an embodiment of the present disclosure, an exemplary volumetric data unfolding method allows a user to inspect the whole inner colon surface on a single 2D image.

Note that while virtual colonoscopy is used as an example for describing embodiments of the present disclosure, the teachings of the disclosure may be used for any virtual endoscopy application.

Referring to FIG. 1, given volumetric data, air inside the colon, soft tissue, and tagged colonic materials are conservatively identified using a histogram based intensity classification 101. These elements, e.g., air inside the colon, soft tissue, and tagged colonic materials, may be classified in the volumetric data. In the application,

"conservatively" means that voxels that may belong to more than one element according to a given metric or expert opinion are classified as UNKNOWN. Consider the case of overlapping soft tissue and tagged colonic material at some intensity range. A voxel in this range may be a soft tissue or tagged material, and thus may be tagged UNKNOWN.

Each voxel is labeled as air, soft tissue, or tagged colonic material. The colon lumen is roughly segmented for centerline extraction 102. It should be noted that not all voxels may be classified and that the segmentation result is an approximation of the colon lumen. An accurate segmentation of colon lumen, in which all voxels are classified, is not necessary. That is, the subsequent colon unfolding result may not rely on this

segmentation result, therefore, an accurate segmentation may not be needed.

A centerline of the colon is extracted from the segmented colon lumen 103. The centerline may be used to provide camera position and orientation for inner colon surface sampling. Any centerline extraction method may be used. The colon unfolding result may not rely on the centerline extraction.

The centerline is uniformly re-sampled and at each sampling position, and a number of rays are cast to detect the inner colon surface 104. A number of points on the centerline may be chosen to cast rays. Here uniformly means the distance between neighboring points is same. Each ray stops at the inner colon surface and returns a 3D position of a last sampling point. Each ray will pass tagged material and stop when it hits the colon surface by using multiple sampling points along the ray.

The sampling points may then be used to fit a surface with B-splines 105, by which the topological noise can be filtered. These sampling points obtained from the ray casting are assumed to be on the colon surface. With these 3D points, a B-spline surface can be fitted (or determined), wherein all the sampling points are on the B-spline surface/colon surface.

Features, such as shape index and curvedness, may be determined using the surface determined by the B-splines (the B-spline surface) 106. The feature information may be used to determine which part of the colon surface is to be preserved during optimization and may not be used for polyp detection. Boundary conditions and constraints based on the calculated features are placed on the fitted colon surface.

A harmonic function may be determined on the colon surface, which may be used to determine 2D coordinates for all vertices of the colon surface 107. As a result, the colon surface is mapped to a planar surface, which can be displayed as a single 2D image and inspected by the physicians.

According to an embodiment of the present disclosure, the ray casting 104 may be combined with surface fitting 105 to generate an inner colon surface and/or to solve a topological noise problem at block 108 as shown in FIG. IB. That is, the fitted colon surface has a B-spline representation, on which features can be determined. The topological noise problem may be an isosurface correction method to detect and remove handles from a mesh representing the colon surface.

In another embodiment of the present disclosure, the processing and rendering can be executed on a multi-core machine where an additional data transfer between a central processor unit (CPU) and a graphics processor unit (GPU) is avoided.

In one embodiment, the colon wall can be extracted accurately and a

segmentation mask can be used in the ray casting 104 for inner colon surface sampling. The segmentation mask may be a binary volume having the same size as the original data. Voxels of a segmentation result, e.g., the inner colon surface determined by the ray casting 104, may be characterized as belonging to the segmentation mask or background voxels.

In one embodiment, features can be calculated from the original CT data instead of the fitted colon surface as shown in FIG. 1C.

It should be understood that exemplary embodiments of the present disclosure are described in terms of CT data, but that the present disclosure is not limited thereto.

Embodiments of the present disclosure may be implemented to other types of volumetric data, including X-ray data, Magnetic Resonance Imaging (MRI) data, etc. It is to be understood that embodiments of the present disclosure may be implemented in various forms of hardware, software, firmware, special purpose processors, or a combination thereof. In one embodiment, a software application program is tangibly embodied on a non-transitory computer-readable storage medium, such as a program storage device or computer program product, with an executable program stored thereon. The application program may be uploaded to, and executed by, a machine comprising any suitable architecture.

According to an embodiment of the present disclosure, a computer system (block 201) for performing a volumetric data unfolding method includes, inter alia, a CPU (block 202), a memory (block 203) and an input/output (I/O) interface (block 204). The computer system (block 201) is generally coupled through the I/O interface (block 204) to a display (block 205) and various input devices (block 206) such as a mouse, keyboard and a slide-scanning microscopy X-Y stage. The support circuits can include circuits such as cache, power supplies, clock circuits, and a communications bus. The memory (block 203) can include random access memory (RAM), read only memory (ROM), disk drive, tape drive, etc., or a combination thereof. The present invention can be

implemented as a module (block 207) of the CPU or a routine stored in memory (block 203) and executed by the CPU (block 202) to process input data (block 208). For example, the data may include image information from a camera, which may be stored to memory (block 403) As such, the computer system (block 201) is a general purpose computer system that becomes a specific purpose computer system when executing the routine of the present disclosure.

The computer platform (block 201) may include a GPU 209 for processing the image data 208. The GPU 209 may be part of a graphics card 210 with dedicated memory 211.

The computer platform (block 201) also includes an operating system and micro instruction code. The various processes and functions described herein may either be part of the micro instruction code or part of the application program (or a combination thereof) which is executed via the operating system. In addition, various other peripheral devices may be connected to the computer platform such as an additional data storage device and a printing device.

It is to be further understood that, because some of the constituent system components and method steps depicted in the accompanying figures may be implemented in software, the actual connections between the system components (or the process steps) may differ depending upon the manner in which the system is programmed. Given the teachings of the present disclosure provided herein, one of ordinary skill in the related art will be able to contemplate these and similar implementations or configurations of the present disclosure.

Having described embodiments for a volumetric data unfolding method, it is noted that modifications and variations can be made by persons skilled in the art in light of the above teachings. It is therefore to be understood that changes may be made in embodiments of the present disclosure that are within the scope and spirit thereof.

Claims

CLAIMS What is claimed is:

1. A method of unfolding volumetric data comprising:

applying a histogram based intensity classification to a volumetric data of a colon for identifying the colon and a material in the colon;

segmenting the colon from the volumetric data;

determining a centerline of the colon;

casting a plurality of rays from the centerline to determine a surface of the colon, wherein the plurality of rays ignore the material in the colon;

resampling the centerline at a plurality of sampling points along the plurality of rays;

fitting the surface of the colon with B-splines using the sampling points to determine a B-spline surface, wherein topological noise is filtered;

determining a feature of the B-spline surface to be preserved; and

unfolding surface of the colon to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B-spline surface and mapping the surface of the B-spline surface to a planar surface, wherein an unfolded surface of the colon is displayed as a two-dimensional image.

2. The method of claim 1, wherein the ray casting is combined with the surface fitting to generate an inner colon surface.

3. The method of claim 1, wherein the ray casting is combined with the surface fitting to determine a topological noise feature.

4. The method of claim 1, further comprising:

labeling every voxel of the surface of the colon; and applying a segmentation mask in the ray casting for inner surface sampling of the colon.

5. A computer program product embodying instructions executed by a processor to perform a volumetric data unfolding method, the method comprising:

applying a histogram based intensity classification to a volumetric data including an object of interest for identifying the object of interest and a material in the object of interest;

segmenting the object of interest from the volumetric data;

determining a centerline of the object of interest;

casting a plurality of rays from the centerline to determine a surface of the object of interest, wherein the plurality of rays ignore the material in the object of interest; resampling the centerline at a plurality of sampling points along the plurality of rays;

fitting the surface of the object of interest with B-splines using the sampling points to determine a B-spline surface, wherein topological noise is filtered;

determining a feature of the B-spline surface to be preserved; and

unfolding surface of the object of interest to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B- spline surface and mapping the surface of the B-spline surface to a planar surface.

6. The computer program product of claim 5, wherein an unfolded surface of the object of interest is displayed as a two-dimensional image.

7. The computer program product of claim 5, wherein the ray casting is combined with the surface fitting to generate an inner surface of the object of interest.

8. The computer program product of claim 5, wherein the ray casting is combined with the surface fitting to determine a topological noise feature.

9. The computer program product of claim 5, further comprising:

labeling every voxel of the surface of the object of interest; and

applying a segmentation mask in the ray casting for inner surface sampling of the object of interest.

10. A computer program product embodying instructions executed by a processor to perform a volumetric data unfolding method, the method comprising:

segmenting the object of interest from the volumetric data;

determining a centerline of the object of interest;

casting a plurality of rays from the centerline to determine a surface of the object of interest, wherein the plurality of rays ignore the material in the object of interest;

determining a feature of the surface to be preserved; and

11. The computer program product of claim 10, wherein an unfolded surface of the object of interest is displayed as a two-dimensional image.

12. The computer program product of claim 10, wherein the ray casting is combined with the surface fitting to generate an inner surface of the object of interest.

13. The computer program product of claim 10, wherein the ray casting is combined with the surface fitting to determine a topological noise feature.

14. The computer program product of claim 10, further comprising:

labeling every voxel of the surface of the object of interest; and

15. A system for performing a volumetric data unfolding method, the system comprising: a processor configured to cast a plurality of rays from a centerline of an object of interest to determine a surface of the object of interest, wherein the plurality of rays ignore at least one type of tagged material in the object of interest,

the processor further configured to resample the centerline at a plurality of sampling points along the plurality of rays,

the processor further configured fit the surface of the object of interest with IB- splines using the sampling points to determine a B-spline surface, wherein topological noise is filtered,

the processor further configured determine a feature of the B-spline surface to be preserved, and

the processor further configured unfold surface of the object of interest to determine a planar surface while preserving the feature by determining coordinates for a plurality of vertices of the B-spline surface and mapping the surface of the B-spline surface to a planar surface; and

a memory configured to store the unfolded surface.