Image analysis – Applications – Biomedical applications
Reexamination Certificate
1998-02-23
2001-08-28
Bella, Matthew C. (Department: 2621)
Image analysis
Applications
Biomedical applications
C382S128000
Reexamination Certificate
active
06282307
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates generally to a method and system for the computerized, automatic delineation of the lung fields in chest radiographs. Specific application is given for the delineation of the costophrenic angles in digitized chest radiographs. Novel developments and implementations include techniques for delineation and splicing of the costophrenic angles with the segmented lung, and improvements in lung segmentation and the assessment of abnormal asymmetry.
The present invention also relates to CAD techniques for automated detection of abnormalities in digital images, for example as disclosed in one or more of U.S. Pat. Nos. 4,839,807; 4,841,555; 4,851,984; 4,875,165; 4,907,156; 4,918,534; 5,072,384; 5,133,020; 5,150,292; 5,224,177; 5,289,374; 5,319,549; 5,343,390; 5,359,513; 5,452,367; 5,463,548; 5,491,627; 5,537,485; 5,598,481; 5,622,171; 5,638,458; 5,657,362; 5,666,434; 5,673,332; 5,668,888; 6,088,473 as well as U.S. application Ser. Nos. 08/158,388; 08,173,935; 08/220,917; 08/398,307; 08/428,867; 08/523,210; 08/536,149; 08/536,450; 08/515,798; 08/562,188; 08/562,087; 08/757,611; 08/758,438; 08/900,191; 08/900,361; 08/900,362; 08/900,188; 08/900,192; 08/900,189; 08/979,623; 08/979,639; 08/982,282; and 09/027,468, each of which are incorporated herein by reference in their entirety.
The present invention also relates to technologies referenced and described in the references identified in the appended APPENDIX and cross-referenced throughout the specification by reference to the number, in brackets, of the respective reference listed in the APPENDIX, the entire contents of which, including the related patents and applications listed above and references listed in the APPENDIX, are incorporated herein by reference.
2. Discussion of the Background
The utility of image processing techniques in diagnostic radiology of the chest has become more pronounced with the growing acceptance of digital radiography, including both direct-digital acquisition and conventional film acquisition with subsequent digitization [
1
]. Techniques for image enhancement such as density correction and unsharp masking [
2
] have, for example, been used to reduce quality variations in portable chest radiographs and to reduce the number of repeat examinations required due to exposure errors [
3
]. Image compression, image transfer protocols, intelligent long-term storage techniques, and interactive display consoles are currently being developed for use with picture archiving and communication systems (PACS) [
4
].
Various image processing methods are being assimilated into computer-aided diagnostic (CAD) schemes [
5
]. Such schemes have been developed for the detection of lung nodules [
6
-
11
], interstitial infiltrates [
8
,
12
-
14
,
11
], pneumothoraces [
15
], cardiomegaly [
16
,
17
], and interval change [
18
].
Inherent in all these schemes is an underlying knowledge of the lung field location in the digital chest radiograph. This has been achieved through the automated detection of intercostal spaces [
19
], rib borders [
20
-
22
], the ribcage edge [
23
], and the complete lung boundary [
24
,
25
]. To detect intercostal spaces, Powell et al. utilized vertical gray-level profiles, to which shift-variant sinusoidal functions were fit [
19
]. Sanada et al. employed a similar method to detect posterior rib borders [
21
]. Statistical analysis of edge gradients and their orientations was then performed within small regions-of-interest (ROIs) to detect subtle continuous rib edges. Wechsler and Sklansky fit linear, parabolic, and elliptical curve segments to the output of gradient and threshold operators to delineate the boundaries of anterior and posterior ribs [
20
]. Chen et al. used edge gradient analysis to determine whether ROIs used for lung texture analysis overlapped rib edges [
22
]. Xu and Doi analyzed the first and second derivatives of gray-level profiles to delineate the ribcage edge [
23
]. Polynomial functions were then fit to initially detected edges. Cheng and Goldberg applied a clustering algorithm to the gray-level histogram computed from a selected region of an image to identify a single gray-level threshold for lung segmentation. The resulting borders were then refined using linear and parabolic curve-fitting techniques. Pietka delineated lung borders using a single threshold determined from the gray-level histogram of a selected region [
25
]. Gradient analysis was then employed to extend the edges.
Others have directly addressed the segmentation of lung fields for the detection of abnormal asymmetry [
26
], for the development of radiographic equalization techniques [
27
], or for use with region-specific display enhancement techniques [
28
-
30
]. Duryea and Boone devised a lung segmentation method based on gray-level profiles and contrast information [
27
]. To selectively enhance the mediastinum and subdiaphragm, Sherrier and Johnson applied histogram equalization techniques to areas determined through local gray-level histogram analysis to be within these regions [
28
]. Sezan et al. identified a lung/mediastinum threshold in the gray-level histogram to perform adaptive unsharp masking in these different anatomic regions [
29
]. McNitt-Gray et al. developed a pattern classification scheme implementing stepwise discriminant analysis as a basis for feature selection, which was then used to train classifiers [
30
]. Clearly, automated segmentation of the lung fields in chest images has many practical applications in addition to its role as a foundation for various CAD schemes.
With the exception of interval change detection, most CAD schemes currently being developed for digital chest radiography are specific to one particular pathology. These schemes often utilize a priori information regarding the “normal” appearance of the ribcage, diaphragm, and mediastinum in a chest image. A potential problem arises when the nature of the thoracic abnormality is such that it substantially affects the volume of the lungs. A large-scale abnormality of this type will usually cause abnormal asymmetry on the radiograph due to a substantial decrease in the area of the aerated lung field (i.e., the high optical density region associated with the normally low attenuation of the lungs) in one hemithorax as projected onto the radiograph. This can substantially alter the overall morphology of the thorax, which, while apparent to a radiologist, could result in the failure of computerized schemes. Such abnormalities would include dense infiltrates, substantial pleural effusions, large neoplasms, extensive atelectasis, pneumonectomy, elevated hemidiaphragm, or cardiomegaly.
A normal PA chest radiograph acquired with the patient properly positioned demonstrates two well-defined CP angles, which represent radiographic projections of the costodiaphragmatic recesses. The costal and diaphragmatic aspects of the normal CP angle converge to form a typically sharp, acute angle. Any observed deviations from this configuration may provide the radiologist with important diagnostic information.
A variety of physical and pathologic conditions may be manifested in the CP angle. In the upright chest examination, for example, non-loculated fluid in the pleural space will collect under the influence of gravity in the costodiaphragmatic recess. Such a pleural effusion will radiographically alter the appearance of the CP angle by blunting the normally sharp appearance of the anatomic recess [
31
,
32
]. In another example, the characteristic flattening of the diaphragm present in patients with emphysema will typically extend into the CP angle region, causing the costal and diaphragmatic aspects of the CP angle to converge at a less acute angle [
31
,
34
]. In addition, fibroti
Armato III Samuel G.
Giger Maryellen L.
MacMahon Heber
ARCH Development Corporation
Bella Matthew C.
Oblon & Spivak, McClelland, Maier & Neustadt P.C.
LandOfFree
Method and system for the automated delineation of lung... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for the automated delineation of lung..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for the automated delineation of lung... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2507707