Добавил:
Sekretar
kiopkiopkiop18@yandex.ru
t.me/Prokururor I Вовсе не секретарь, но почту проверяю
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз:
Предмет:
Файл:Ординатура / Хирургия / Библиотека им академика М.И. Перельмана / Книга_5829_Библиотеки_им_академика_М_И_Перельмана.pdf
Media and Intima Thickness and Texture Analysis of the Common Carotid Artery 125
Credit/copyright notice: Based on “Manual and automated media and intima
thickness measurements of the common carotid artery,” by Loizou CP, Pattichis CS,
Nicolaides A, Pantziaris M (2009) IEEE Trans Ultrason Ferroelectr Freq Control
56(5):983–994,c 2009EEE, and on “Ultrasound image texture analysis of the
intima and media layers of the common carotid artery and its correlation with age
and gender” by Loizou CP, Pantziaris M, Pattichis MS, Kyriakou E, Pattichis CS
(2009) Comput Med Imaging Graph 33(4):317–324,c CMIG2009.
References
1. American Heart Association (2007) Heart disease and stroke statistics-2008, update, Dallas.
http://www.americanheart.org/presenter.jhtml.
2. Pignoli P, Tremoli E, Poli A, Oreste P, Paoleti R (1986) Intima plus media thickness of the
arterial wall: a direct measurement with ultrasound imaging. Atherosclerosis 74(6):1399–1406
3. Touboul P-J, Labreuche J, Vicaut E, Amarenco P (2005) Carotid intima–media thickness,
plaques, and Framingham risk score as independent determinants of stroke risk. Stroke
36(8):1741–1745
4. Watanabe T, Koba S, Kawamura M, Itokawa M, Idei T, Nakagawa Y, Iguchi T, Katagiri T
(2004) Small dense low-density lipoprotein and carotid atherosclerosis in relation to vascular
dementia. Metabolism 53(4):476–482
5. Mario CD, Gorge G, Peters R, Pinto F, Hausmann D, von Birgelen C, Colombo A, Murda H,
Roelandt J, Erbel R (1998) Clinical application and image interpretation in coronary ultrasound. Study group of intra-coronary imaging of the working group of coronary circulation
and of the subgroup of intravascular ultrasound of the working group of echocardiography of
the European Society of Cardiology. Eur Heart J 19(2):201–229
6. Grønhold ML, Nordestgaard BG, Schroeder TV, Vorstrup S, Sillensen H (2001) Ultrasonic
echolucent carotid plaques predict future strokes Circulation 104(1):68–73
7. Wilhjelm JE, Grønholdt MLM, Wiebe B, Jespersen SK, Hansen LK, Sillensen H (1998)
Quantitative analysis of ultrasound B-mode images of carotid atherosclerotic plaque: correlation with visual classification and histological examination. IEEE Trans Med Imaging
17(6):910–922
8. Gussenhoven EJ, Frietman PA, van Suylen SHRJ, van Egmond FC, Lancee CT, van Urk H,
Roelandt JR, Stijnen T, Bom N (1991) Assessment of medial thinning in atherosclerosis by
intravascular ultrasound Am J Cardiol 68:1625–1632
9. Wendelhag I, Liang Q, Gustavsson T, Wikstrand J (1997) A new automated computerized
analysing system simplifies reading and reduces the variability in ultrasound measurement of
intima media thickness. Stroke 28:2195–2200
10. Loizou CP, Pattichis CS, Pantziaris MS, Tyllis T, Nicolaides AN (2007) Snakes based
segmentation of the common carotid artery intima media. Med Biol Eng Comput 45:35–49
11. Loizou CP, Pantziaris M, Pattichis MS, Kyriakou E, Pattichis CS (2009) Ultrasound image
texture analysis of the intima and media layers of the common carotid artery and its correlation
with age and gender. Comput Med Imaging Graph 33(4):317–324
12. Loizou CP, Pattichis CS, Nicolaides A, Pantziaris M (2009) Manual and automated media and
intima thickness measurements of the common carotid artery. IEEE Trans Ultrason Ferroelectr
Freq Control 56(5):983–994
13. Gutierrez M, Pilon P, Lage S, Kopel L, Carvalho R, Furuie S (2002) Automatic measurement
of carotid diameter and wall thickness in ultrasound images. Comput Cardiol 29:359–362
14. Delsanto S, Molinari F, Giustetto P, Liboni W, Badalamenti S, Suri JS (2007) Characterization
of a completely user independent algorithm for carotid artery segmentation in 2-D ultrasound
images. IEEE Trans Instrum Meas 56(4):1265–1274
126 C.P. Loizou et al.
15. Rodriguez-Maciasa KA, Lindbc L, Naessena T (2006) Thicker carotid intima layer and thinner
media layer in subjects with cardiovascular diseases: an investigation using noninvasive highfrequency ultrasound. Atherosclerosis 189(2):393–400
16. Osika W, Dangardt F, Gr¨onros J, Lundstam U, Myredal A, Johansson M, Volkmann R,
Gustavsson T, Gan LM, Friberg P (2007) Increasing peripheral artery intima thickness from
childhood to seniority. Arterioscler Thromb Vasc Biol 27:671–676
17. Mintz GS, Nissen SE, Anderson W, Bailey SR, Erbel R et al (2001) American college of
cardiology clinical expert consensus document on standards for acquisition, measurements and
reporting intravascular ultrasound studies (IVUS). J Am Coll Cardiol 37:1478–1492
18. Loizou CP, Pattichis CS, Pantziaris M, Nicolaides A, Georgiou N, KyriakouE (2007) Media
thickness measurement of the common carotid artery. Proceedings of the 29th annual international conference of the IEEE engineering in medicine and biology society, Cite Int., Lyon,
FrP1B6.5, 23–26 Aug 2007
19. Kyriakou E, Pattichis MS, Christodoulou CH, Pattichis CS, Kakkos S, Griffin M, Nicolaides
AN (2005) Ultrasound imaging in the analysis of carotid plaque morphology for the assessment
of stroke. In: Suri JS, Yuan C, Wilson DL, Laxminarayan S (eds) Plaque imaging: pixel to
molecular level, IOS, pp 241–275
20. Tegos TJ, Sabetai MM, Nicolaides AN, Elatrozy TS, Dhanjil S, Stevens JM (2001) Patterns
of brain computed tomography infraction and carotid plaque echogenicity. J Vasc Surg
33:334–339
21. Williams DJ, Shah M (1992) A fast algorithm for active contour and curvature estimation.
Comput Vis Image Underst 55(1):4–26
22. Chalana V, Kim Y(1997) A methodology for evaluation of boundary detection algorithms on
medical images. IEEE Trans Med Imaging 16(5):642–652
23. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two
methods of clinical measurement. Lancet 1(8476):307–310
24. Mojsilovic A, Popovic M, Amodaj N, Babic R, Ostojic M (1997) Automatic segmentation of
intravascular ultrasound images: a texture based approach. Ann Biomed Eng 25:1059–1071
25. Loizou CP, Pattichis CS, Pantziaris MS, Tyllis T, Nicolaides AN (2006) Quantitative quality
evaluation of ultrasound imaging in the carotid artery. Med Biol Eng Comput 44(5):414–426
26. Loizou CP, Pattichis CS, Pantziaris MS, Nicolaides AN (2007) An integrated system for the
segmentation of atherosclerotic carotid plaque. IEEE Trans Inf Technol Biomed 11(5):661–667
27. Nicolaides AN, Kakkos SK, Griffin M, Sabetai M, Dhanjil S, Thomas D et al (2005) Effect of
image normalization on carotid plaque classification and the risk of ipsilateral hemispheric
events: results from the asymptomatic carotid stenosis and risk of stroke study. Vascular
1(4):211–221
28. Graf S, Gariery J, Massonneau M, Armentano R, Mansour S, Barra J, Simon A, Levenson
J (1999) Experimental and clinical validation of arterial diameter waveform and intimal
media thickness obtained from B-mode ultrasound image processing Ultrasound Med Biol
25(9):1353–1363
29. Mancini GBJ, Abbott D, Kamimura C, Yeoh E (2004) Validation of a new ultrasound method
for the measurement of carotid artery intima medial thickness and plaque dimensions. Can J
Cardiol 20(13)1355–1359
30. Christodoulou CI, Pattichis CS, Pantziaris MS, Nicolaides AN (2003) Texture-based classification of atherosclerotic carotid plaques. IEEE Trans Med Imaging 22(7):902–912
31. Belcaro G, Nicolaides AN, Laurora G, Cesarone MR, De Sanctis M, Incandela L, Barsotti A
(1996) Ultrasound morphology classification of the arterial wall and cardiovascular events in a
6-year follow-up study. Arterioscl Thromb Vasc Biol 16:851–856
32. Schmidt C, Wendelhag I (1999) How can the variability in ultrasound measurements of intima–
media thickness be reduced? Studies of interobserver variability incarotid and femoral arteries.
Clinic Physiol 19(1):45–55
33. Ellis SM, Sidhu PS (2000) Granularity of the carotid artery intima–medial layer: reproducibility of quantification by a computer-based program. Br J Radiol 37:595–600
Media and Intima Thickness and Texture Analysis of the Common Carotid Artery 127
34. Bartolomucci F, Paterni M, Morizzo C, Kozakova M et al (2001) Early structural changes of
carotid artery in familial hypercholesterolemia. J Clin Hypertens 14:125A–126A
35. Lind L, Andersson J, Roenn M, Gustavsson T (2007) The echogenecity of the intima–media
complex in the common carotid artery is closely related to the echogenecity in plaques.
Atherosclerosis 195:411–414
36. (2004) Prevention of disampling and fatal strokes by successful carotid endarterectomy in patients without recent neurological symptoms: randomized control trial, Lancet
363(9420):1491–1502
37. Rosfors S, Hallerstam S, Jensen-Urstad K, Zetterling M, Carlstroem C (1998) Relationship
between intima–media thickness in thecommon carotid artery and atherosclerosis in the carotid
bifurcation. Stroke 29:1378–1382
38. Balasundaram JK, Wahida Banu RSD (2006) A non-invasive study of alterations of the carotid
artery with age using ultrasound images. Med Biol Eng Comput 44:67–72
39. Haralick RM, Shanmugam K, Dinstein I (1973) Texture features for image classification. IEEE
Trans Syst Man Cybern SMC 3:610–621
40. Weszka JS, Dyer CR, Rosenfield A (1976) A comparative study of texture measures for terrain
classification. IEEE Trans Syst Man Cybern SMC 6:269–285
41. Amadasun M, King R (1989) Textural features corresponding to textural properties. IEEE
Trans Syst Man Cybern 19(5):1264–1274
42. Wu CM, Chen YC, Hsieh K-S (1992) Texture features for classification of ultrasonic images.
IEEE Trans Med Imaging 11:141–152
43. Litwin M, Wuehl E, Jourdan C, Trelewicz J, Niemirska A, Fahr K, Jobs K, Grena R, Wawer
ZT, Rajszys P, Troeger J, Mehls O, Shaefer F (2005) Altered morphological properties of large
arteries in children with chronic renal failure and after and after renal transplantation. J Am
Soc Nephrol 16:1494–1500
44. Goncalves I,Lindholm MW, Pedro LM et al (2004) Elastin and calcium rather than collagen or
lipid content are associated with echogenecity of human carotid plaques. Stroke 35:795–800
Christos P. Loizou was born in Cyprus
in October 1962 and received his BSc
degree in Electrical Engineering, the
Dipl.-Ing. (MSc) degree in Computer Science and Telecommunications from the
University of Kaisserslautern, Kaisserslautern, Germany, and the PhD degree
from the Department of Computer Science, Kingston University, London, UK,
on ultrasound image analysis of the
carotid artery in 1990 and 2005, respectively. He is currently an Assistant Professor with the Department of Computer
Science at Intercollege in Cyprus. His
research interests include medical imaging, in investigating the risk of stroke and
the multiple sclerosis disease.
128 C.P. Loizou et al.
Marios Pantziaris received the M.D.
degree in neurology from the Aristote-
lion University, Thessaloniki, Greece, in
1995. Currently, he is working with
Cyprus Institute of Neurology and Ge-
netics, Nicosia, Cyprus, as a Senior Neu-
rologist in the Neurological Department
and is the Head of the Neurovascu-
lar Department. He has been trained
in Carotid Duplex–Doppler ultrasonogra-
phy at St. Mary’s Hospital, London, in
1995. In 1999, he was a visiting doctor
in acutestroke treatmentat Massachusetts
General Hospital, Harvard University,
Boston. He has considerable experience in carotids–transcranial ultrasound, has
participated in many research projects, and has several publications to his name. He
is also the Head of the Multiple Sclerosis (MS) Clinic where he is running research
projects towards the aetiology and therapy of MS.
Constantinos S. Pattichis was born in
Cyprus on January 30, 1959, and received
his diploma as a technician engineer from
the Higher Technical Institute in Cyprus
in 1979, BSc in Electrical Engineering
from the University of New Brunswick,
Canada, in 1983, MSc in Biomedical En-
gineering from the University of Texas at
Austin, USA, in 1984, MSc in Neurology
from the University of Newcastle Upon
Tyne, UK, in 1991, and PhD in Elec-
tronic Engineering from the University
of London, UK, in 1992. He is currently
Professor with the Department of Computer Science of the University of Cyprus.
His research interests include ehealth, medical imaging, biosignal analysis, and
intelligent systems.
CAUDLES-EF: Carotid Automated
Ultrasound Double Line Extraction
System Using Edge Flow
Filippo Molinari, Kristen M. Meiburger, Guang Zeng, Andrew Nicolaides,
and Jasjit S. Suri
Abstract The evaluation of the carotid artery wall is essential for the diagnosis of
cardiovascular pathologies or for the assessment of a patient’s cardiovascular risk.
This chapter presents a complete user-independent algorithm, which automatically extracts the far double line (lumen–intima and media–adventitia)in the carotid
artery using an Edge Flow technique based on the directional probability maps
using the attributes of intensity and texture. Specifically, the algorithm traces the
boundaries between the lumen and intima layer (line one) and between the media
and adventitia layer (line two). The Carotid Automated Double Line Extraction
System based on Edge Flow (CAUDLES-EF) ischaracterized and validated by comparing the output of the algorithm with the manual tracing boundariescarried out by
three experts. We also benchmark our new technique with the two other completely
automatic techniques (CALEXia and CULEXsa) that we had previously published.
Our multi-institutional database consisted of 300 longitudinal B-mode carotid
images with normal and pathologic arteries. We compared our current new method
F. Molinari () • K.M. Meiburger
Biolab, Department of Electronics, Politecnico di Torino, Corso Duca degli Abruzzi,
24, 10129 Torino, Italy
e-mail: filippo.molinari@polito.it; kristen.meiburger@polito.it
G. Zeng
MBF Bioscience Inc. Williston, VT, USA
e-mail: gzeng@clemson.edu
A. Nicolaides
Vascular Screening and Diagnostic Centre, London, UK
Department of Biological Sciences, University of Cyprus, Nicosia, Cyprus
e-mail: anicolaides1@gmail.com
J.S. Suri
Biomedical Technologies, Inc., Denver, CO, USA
Idaho State University (Affiliated), Pocatello, ID, USA
e-mail: jsuri@comcast.net
J.M. Sanches et al. (eds.), Ultrasound Imaging: Advances and Applications,
DOI 10.1007/978-1-4614-1180-2
6, © Springer Science+Business Media, LLC 2012
129
130 F. Molinari et al.
with previous methods, and showed the mean and standard deviation for the
three methods: CALEXia, CULEXsa, and CAUDLES-EF as: 0.134 ±0.088mm,
0.74 ±0.092mm and 0.043±0.097mm, respectively. Our IMT was slightly underestimated with respect to the ground truth of the IMT, but showed a uniform
behavior over the entire database. As in view of regards the Figure of Merit (FoM)
for CALEXia and CULEXsa showed the values of 84.7%, and 91.5%, respectively,
while our new approach, CAUDLES-EF performed the best at 94.8%, showing a
good improvement compared to previous methods.
Keywords Carotid artery • Ultrasound • Multi-resolution • Edge Flow • Localization • Intima–media thickness • Hausdorff distance • Polyline distance
1 Background
Numerous studies from around the world have demonstrated that there is a strong
correlation between the risk of cerebrovascular diseases and the characteristics of
the carotid artery wall [1–3]. Ultrasound examination is a widely used diagnostic
tool for assessing and monitoring the plaque build up via the carotid window.
Ultrasounds offer several advantages in clinical practice:
1. They only propagate mechanical (i.e., non-ionizing) radiation.
2. No short-term or long-term adverse biological effects have been demonstrated in
the power and intensity range commonly used in clinical scans.
3. The examination is quick and safe.
4. The ultrasound equipment is one among the more inexpensive equipment when
compared to other imaging devices.
However, ultrasound examinations are operator-dependent and the ultrasound im-
ages can tend to be quite noisy and require training for correct interpretation.
The intima–media thickness (IMT) is the most used and validated marker of
progression of carotid artery diseases [4, 5] and can be measured using image
processing strategies and ad-hoc computer techniques. The goal is to first segment
the carotid artery distal wall, so as to then find the lumen–intima (LI) and media–
adventitia (MA) boundaries. The distance calculated between these two interfaces
is taken as an estimate of the IMT.
The segmentation process can conceptually consist of two cascading stages:
• Stage I: recognitionof the carotid artery (CA) and delineation ofthe far adventitia
layer (AD
• Stage II: tracing of the LI/MA wall boundariesin the ROI of the recognized CA.
In Stage I, the carotid artery must be correctly located within the ultrasound image
frame. This stage is generally performed better by human experts, who can mark
the position of the CA by either tracing rectangular regions-of-interest (ROI) or
by placing the markers. In Stage II, the guidance zone is created in which the LI
and MA border are estimated. The IMT can be subsequently measured once the LI
and MA borders are determined during the segmentation process. These two stages
) in the two-dimensional B-mode ultrasound image.
F
CAUDLES-EF: Carotid Automated Ultrasound Double... 131
cannot be independentfrom each other.In fact, StageI is of fundamental importance
since the AD
profile is found and used as a starting point for Stage II, which is
F
also automated. The performance of Stage I will directly affect the initialization of
Stage II, and also affect the final results. This fact emphasizes the importance of the
need of an accurate yet versatile technique to perform Stage I.
In orderto achievecomplete automation,both of these stages mustbe designed to
be independent of the user. To do so, first of all appropriate detection strategies are
required to automatically locate the carotid artery in the image. These strategies
must be robust with respect to noise and must be able to process carotids with
different geometrical appearance.
The majority of the algorithms proposed in the literature for the automated
segmentation of the CA in ultrasound images require a certain degree of userinteraction, which precludes real complete automation. Any user-interaction that
slows down the analysis process would introduce a dependence on the operator
if gain settings are not optimal, bringing subjectivity into the process. Complete
automation, instead, can be an asset for multi-center large studies since it enables
the processing of large image databases.
This chapter presents a complete user-independent Carotid Automated Double
Line Extraction System using Edge Flow (CAUDLES-EF) algorithm, which performs both Stages I and II. Starting from the ultrasound image, the algorithm first
segments the distal border of the CA and then performs the automatic detection of
the LI and MA interfaces. Neither of these processes requires any user-interaction.
The first part of our new algorithm is based on scale-space multi-resolution analysis
while the second part is based on flow field propagation. CAUDLES-EF was
specifically designed for the IMT measurement of the far (distal) wall of the
common carotid artery. We also show the characterization of this algorithm in
terms of automatic versus human traced segmentation, and we also benchmarked
the results with two other completely automatic techniques that we had previously
developed [6–10]. Our image database consisted of 300 images coming from
two different institutions consisting of both normal and pathological arteries. Two
different technicians acquired the images, using two different ultrasound scanners.
We used the Hausdorff distance as a performance metric for Stage I and we
measured the distance between the computed far adventitial wall and the LI/MA
profiles that were manually traced by experts, the so-called ground truth. For
assessing the performance of Stage II, we used the Polyline distance and calculated
the error between the IMT estimated using our algorithm and ground truth IMT.
2 Materials and Methods
2.1 Image Database and Preprocessing Steps
We tested an image database consisting of 300 images coming from two different
Institutions. Two hundred images were acquired using an ATL HDI 5000 ultrasound
132 F. Molinari et al.
scanner equipped by a 10–12 MHz probe at the Neurology Division of Gradenigo
Hospital (Torino, Italy), from one hundred and fifty asymptomatic patients who
referred to the Neurology Division for carotid assessment (age: 69 ±16 years old;
range: 50–83 years old). Ninety-three subjects were male. Of these subjects, 80 had
hypertension history, 40 had hypercholesterolemia, and 30 had both. Ten patients
were diabetic. Resampling was set to 16 pixels/mm, leading to an axial resolution
equal to 62.5 μm/pixel. The remaining one hundred images were acquired at The
Cyprus Institute of Neurology and Genetics (Nicosia, Cyprus) from asymptomatic
individuals (age: 54 ±24; range: 25–95 years) using a Philips ATL HDI 3000
ultrasound scanner equipped with a linear 7–10 MHz probe. These images were
resampled at a density of 16.67 pixels/mm, therefore obtaining an axial spatial
resolution equal to 60 μm/pixel. Both of the Institutions made sure to obtain written
informed consent from the patients prior to enrolling them in the study and also
got approval by the respective IRBs. Both the experimental protocol and the data
acquisition procedure were approved by the respective local Ethical Committees.
Three different expert sonographers (a cardiologist, a vascular surgeon, and a
neurologist – all with more than 20 years of experience in their field) independently
manually segmented the images by tracing the boundariesof the lumen–intima (LI)
and media–adventitia (MA) interfaces, and the average tracings were considered as
ground truth (GT).
In order to discard the surrounding black frame containing device headers
and image/patient test data, the raw ultrasound image is automatically cropped
in one of two ways. The first method is for DICOM images with fully formatted DICOM tags: we used the data contained in the specific field named
SequenceOfUltrasoundRegions, which contains four subfields that mark
the location of the image which contains the ultrasound representation. These fields
are named RegionLocation (with their specific labels being: x
and y
) and they mark the horizontal and vertical extensions of the image. The
max
min
, x
max
, y
min
raw B-Mode DICOM image is then cropped in order to extract only the portion
which contains the carotid morphology.If, however, the image was not in a DICOM
format or if the DICOM tags were not fully formatted, the second method was
applied: adopting a gradient-based procedure, we computed the horizontal and
vertical Sobel gradients of the image. When computed outside of the region of
the image containing the ultrasound data, gradients are equal to zero. Hence, the
beginning of the image region containing the ultrasound data can be found as the
first row/column with a gradient different from zero. Similarly, the last non-zero
row/column of the gradient marks the end of the ultrasound region.
,
2.2 Architecture of CAUDLES-EF
2.2.1 Stage I: Far Adventitia Estimation
Since the CAUDLES-EF algorithm was developed to help in reducing the human
operator dependence and therefore be totally user-independent, the first stage of
CAUDLES-EF: Carotid Automated Ultrasound Double... 133
Fig. 1 CAUDLES-EF procedure for ADFtracing. (A) Original cropped image. (B) Downsampled
image. (C) Despeckled image. (D) Image after convolution with first-order Gaussian derivative
(σ = 8). (E) Intensity profile of the column indicated by the vertical dashed line in panel D. (AD
indicates the position of the far adventitia wall). (F) Cropped image with far adventitia profile
overlaid
F
the algorithm is the completely automatic recognition of the CA. This is done
through a novel and low-complexity procedure, which detects the far adventitia
border using a method based on scale-space multi-resolution analysis. Starting from
the automatically cropped image (Fig. 1a), the automated Stage I is divided into
different steps which is described in detail here:
• Step 1: Fine to Coarse Down-sampling. The image is first down-sampled by a
factor of two (i.e., the number of rows and columns of the image is halved)
(Fig. 1b) and implementing the down-sampling method as discussed by Ye et al.
[11] by adopting a bi-cubic interpolation. This method was tested on ultrasound
images and showed a good accuracy and a low computational cost [11].
• Step 2: Speckle reduction. Speckle noise is attenuated using a first-order local
statistics filter (called lsmv by the authors [12, 13]), which has given the best
performance in the specific case of carotid imaging. This filter is defined by the
following equation:
J
= I + k
x,y
where I
pixel neighborhood, and k
in the moving window is indicated by J
k
x,y
and
is the intensity of the noisy pixel, I is the mean intensity of a N × M
x,y
is a local statistic measure. The noise-free central pixel
x,y
2
σ
I
=
I
2
σ
n
,where
2
2
2
σ
+
σ
n
I
the variance of the noise in the cropped image. An optimal neighborhood
2
σ
represents the variance of the pixels in the neighborhood,
I
x,y
. Louizou et al. [12, 13] mathematically
x,y
I
−I
x,y
(1)
size was demonstrated to be 7 ×7. Figure 1c shows the despeckled image.
134 F. Molinari et al.
• Step 3: Higher order Gaussian derivative filter. The despeckled image is then
filtered using a 35 ×35 pixels first-order derivative of a Gaussian kernel. The
scale parameter of the Gaussian derivative kernel is taken equal to 8 pixels. This
value is chosen because it is equal to half the expected dimension of the IMT
value in an original fine resolution image, since an average IMT value equal
to 1 mm corresponds roughly to about 16 pixels in the original image scale and
therefore 8 pixels in the down-sampled image. The white horizontal stripes in
Fig. 1d show the proximal (near) and distal (far) adventitia layers.
• Step 4: Automated Far Adventitia (AD
)tracing.Figure 1e shows the intensity
F
profile of one column (from the upper edge of the image to the lower edge of the
image) of the Gaussian filtered image. The proximal (near) and distal (far) walls
are clearly identifiable as intensity maxima saturated to the maximum value of
255. A heuristic search is then applied to the intensity profile of each column to
automatically trace the profile of the distal (far) wall. The image convention uses
(0, 0) as the top left-hand corner of the image, and so this search is done starting
from the bottom of the image (i.e., from the pixel with the highest row index)
and searching for the first white region consisting of at least 6 pixels (computed
empirically). The deepest point of this region (i.e., the pixel with the highest row
index) marks the position of the far adventitia AD
overall automated AD
tracing is found as the sequence of points resulting from
F
layer on that column. The
F
the heuristic search for all of the image columns.
• Step 5: Up-sampling of the far adventitia (AD
) boundary locator. The AD
F
profile that is found is then subsequently up-sampled to the original fine scale
and superimposed overthe original cropped image (Fig.1f) for both visualization
and performance evaluation.
F
This Stage I consists essentially of an architecture based on fine-to-coarse sampling
for vessel wall scale reduction, speckle noise removal, higher-order Gaussian
convolution, and automated recognition of the far adventitia border. This multiresolution method prepares the vessel wall’s edge boundary so that the thickness of
the vessel wall is roughly equivalentto the scale of the Gaussian kernels. This allows
an optimal detection of the CA walls since when the kernel is located close to a near
gray level change, it enhances the transition. Consequently, the most echoic image
interfaces are enhanced to white in the filtered image. This fact is clearly shown in
Fig. 1d, e, is what allows this procedure to detect the far adventitia layer.
2.2.2 Stage II: Double Line (LI/MA) Border Estimation
The whole idea of double line extraction for IMT measurement in stage II is to
first extract the strong LI/MA edges which lie between lumen region and AD
border. There are two kind of strong edges: LI strong edges and MA strong edges.
The search for the double line (LI/MA borders) strong edges is performed in the
grayscale guidance zone. This grayscale guidance zone is computed empirically
from the knowledge database. The strong edge estimation in this guidance zone
F