Microscopic Image Analysis Techniques for the

Morphological Characterization of Pharmaceutical Particles:

Influence of Process Variables
SERGIO ALMEIDA-PRIETO, JOSÉ BLANCO-MÉNDEZ, FRANCISCO J. OTERO-ESPINAR
Departamento de Farmacia y Tecnologı́a Farmacéutica, Universidad de Santiago de Compostela,
15076 Santiago de Compostela, La Coruña, Spain

Received 12 May 2005; revised 12 September 2005; accepted 22 September 2005

Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20507

ABSTRACT: This study investigated the effects of various process variables on some of
the shape factors most widely used in the morphological characterization of pharmaceu-
tical particles (including circularity, aspect ratio, er, and the recently proposed Vr and
Vp). After optimization of the illumination system and greyscale thresholds for
discrimination of particle background, we evaluated the effects of process variables
within the image capture and analysis system, including the image capture device (video
or digital camera), image color information (24-bits-per-pixel RGB or 8-bits-per-pixel
black and white), file type (JPG or TIFF), and JPG file compression ratio. A key aspect was
evaluation of the effect of scaling factor (mm/pixel), dependent on real pixel size and
optical magnification, on shape factor estimates. The results obtained indicate that
accurate estimation of shape factors requires use of a scaling factor below a certain
maximum; however, use of very low scaling factors will mean that the field of view is very
small, so that it will contain very few particles, implying a time-consuming increase in the
number of images that must be analyzed. Finally, we use statistical procedures to
estimate the minimum number of particles that must be analyzed in order to provide
accurate estimates of shape factors. ß 2005 Wiley-Liss, Inc. and the American Pharmacists
Association J Pharm Sci 95:348–357, 2006
Keywords: image analysis; granulation; shape analysis; extrusion; spheronization;
pixel size; morphology; pellets

INTRODUCTION paleontology, geology, and metallurgy.1–5 Before

The enormous advances in computing and ima-
ging technologies in recent years have led to
increasing interest in the applications of image
analysis (IA) in diverse fields of research. Analy-
sis of the size and shape of different types of
particle is of great interest in pharmaceutical
technology, and of course in many other fields;
indeed most shape factors and other parameters
used in IA in pharmaceutical technology were
originally developed in other fields, such as
Correspondence to: F.J. Otero-Espinar (Telephone; 34 981
563100 ext. 14878; Fax: 34 981 547148;
E-mail: ffrotero@usc.es)
Journal of Pharmaceutical Sciences, Vol. 95, 348–357 (2006)
ß 2005 Wiley-Liss, Inc. and the American Pharmacists Association
the appearance of IA software, however, dimen-
sions were determined on the basis of direct
measurement under a microscope, and in most
cases the morphological parameters were calcu-
lated by general procedures, often developed in
other fields. This approach is time-consuming and
has poor measurement precision, but the only
significant source of variability is inter-operator
variability. In the newly available IA systems,
based on photomicroscopes or videomicroscopes
linked to personal computers, however, the
potential sources of variability are more diverse.
The influence of various factors on the precision
of computer-based IA methods was first discussed
by Lövgren and Lundberg [1989].6 In 1992
Zingermenn et al.7 published one of the first

348 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006


studies in pharmaceutical technology to point out
that IA systems need to be properly validated for
effective use. This article centered largely on the
validation of aspects related to light microscopy
components of IA systems. Subsequently, in 1995,
Podczeck et al.8 noted that important sources of
variation in the morphological characterization of
particles by IA are illumination intensity and
direction, greyscale thresholding choices, and
pixel size (minimum pixel size should be deter-
mined in view of the optical magnification used).
Over the last decade, however, there have been
dramatic advances in digital photography and in
computing systems for image treatment and
management. These advances have opened up
new and interesting applications for IA techni-
ques, and have led to clear improvements in the
quality of the results obtained. At the same time,
however, there has been an increase in the number
of potential sources of variation. For example,
modern digital cameras can obtain information
with different levels of resolution, and this infor-
mation can be stored in various file types each with
different data compression characteristics, and
consequent effects on image characteristics.
Furthermore, depending on the type of camera
used and above all the size and type of the sensor
used for image capture (CMOS or CDC), pixel size
and quality varies markedly.
In view of these considerations, the principal
aim of the present study was to identify the most
important sources of variation in a number of
parameters commonly used in the morphological
characterization of pharmaceutical particles. Spe-
cifically, we considered two widely used shape
factors, circularity and aspect ratio, as well as er as
defined by Podczeck and Newton [1994] and
Podczeck et al. [1999]8,9 and the parameters Vr
and Vp recently proposed by Almeida et al [2004].11
These parameters are all widely used in proce-
dures using recently developed image capture and
image management technologies.
MATERIALS AND METHODS
All particles used in this study were pellets
prepared by extrusion-spheronization, from dif-
ferent mixtures of starches and dextrins, using
water as wetting agent. With the aim of obtaining
batches with different morphologies, two different
starch/dextrin proportions were used, one giving
largely spherical pellets and the other largely
irregular pellets. The components were mixed in a
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006
MICROSCOPIC IMAGE ANALYSIS TECHNIQUES 349
Turbula T2 mixer for 15 min, then kneaded with
appropriate quantities of water in a Kenwood
Chef Classic orbital mixer for 15 min; then extru-
ded in a Caleva 10 extruder (6 rpm, hole diameter
1 mm); then spheronized in a Caleva 120 sphe-
ronizer (2000 rpm, 15 min); then dried to constant
weight in a Heraeus oven.
Images were obtained with an Olympus SZ-
CTV stereomicroscope with different image-cap-
ture devices attached: (1) a JVC TK-S350 black-
and-white video camera (1/3-inch interline-trans-
fer CCD with effective pixel number of 430000)
connected to a Matrox Comet video card via PC
Image VGA 24 software, 92) an Olympus DP11
microscope camera system, or (3) an Olympus
DP12 microscope camera system. Samples were
illuminated perpendicularly from above, against a
black background, as recommended by Podczeck
and Newton [1995],10 using a Highlight 2000 cold
light (Olympus Europe). A total of 50 particles
from the spherical-pellet batches and 50 particles
from the irregular-pellet batches were studied. To
avoid variations associated with the position of
each particle in the field of view, each particle was
glued to the black background.
Shape factor values were calculated from the
particle measurement data using the SigmaScan
package and various purpose-written macros in
Microsoft Excel XP.
To investigate the influence of file compression
method on shape factors, 20 spherical particles
were measured in sextuplicate using the Olympus
DP12 camera system, which allows the images to
be stored in either TIFF or JPG format, in the
latter case with varying compression ratio.
Shape factors were compared among treat-
ments by two-way analysis of variance following
confirmation of data normality and variance
homogeneity. Means were subsequently compared
by the Student-Newman-Keuls test. All statistical
analyses were performed using the SigmaSat
1.0 statistics package.
RESULTS AND DISCUSSION
Effects of Sample Illumination Procedure
One of the first questions to be taken into account
when validating image analysis systems is the
sample illumination system. Two aspects of
sample illumination are important. The first is
related to the position of the illumination source
with respect to the sample.9,10
350 ALMEIDA-PRIETO, BLANCO-MÉNDEZ, AND OTERO-ESPINAR

The second important aspect of sample illumi-
nation is the intensity of the light source. Figure 1
summarizes the procedure used to determine
optimal illumination intensity in the present
study, taking as an example image capture using
the JVC TK-S350 video camera. As can be seen, the
importance of illumination intensity depended on
the shape factor considered. Thus Vr, AR, and er
were scarcely affected by illumination intensity,
whereas Vp increased markedly at light intensi-
ties above 3 (on the arbitrary scale used by the
Highlight Olympus cold light). If the light source
used is too faint, the contrast between particle and
background will be weak, and identification of the
particle edge will be strongly dependent on the
greyscale threshold choice (or equivalent color-
mapping criteria). In contrast, if the light source is
too strong, the excessive brightness will hinder
focusing and reduce edge definition. This effect on
Vp is basically attributable to the fact that high
light intensities hindered effective measurement
of perimeter but did not markedly affect measure-
ments of other parameters (radius, length, width,
area). On the basis of these results, we conclude
that for our system the best compromise between
brightness is obtained at about light intensity 3,
independently of particle shape.

A Effects of Thresholding
12 Vr
Vp
For measurement of the particle dimensions used
er
10 in morphological analysis, analog images (as
AR
obtained with the video camera in the present
study) must first be converted into a binary image
Vr, Vp, V, er, AR

8
with one bit per pixel. The software must then
6
count the number of pixels making up each
dimension (area, perimeter, etc.) before multi-
plying by the corresponding scaling factor. When
4
the image captured is already digital (i.e., digital
camera or scanner), it must be converted to a bit
2
map comprising a series of pixels whose size will
vary depending on resolution and optical magni-
0 fication. Each of the pixels has an associated color
2.0 2.5 3.0 3.5 4.0 4.5 5.0
value, depending on the color mode being used
Light intensity
(black and white, greyscale, red-green-blue, etc.).
For measurement of particle dimensions, the
image must be converted to a binary image,
B 30
requiring selection of threshold values for con-
Vr
Vp verting greyscale or color images to black-and-
25 er white images with one bit per pixel.
AR Figure 2 shows the results of using different
20
threshold values for analysis of spherical (A) and
irregular particles (B). As with light intensity
Vr, Vp, V, er, AR

(except for its effects on Vp), the choice of threshold

15 value had little effect on measured particle
dimensions such as radius, length and width.
10 Again, the shape factor most markedly affected by
choice of threshold was Vp: threshold values lower
than 55–60% gave higher Vp values. This is
5
attributable to the fact that the threshold value
directly determining which pixels will be included
0 or excluded from the definition of the particle and
2.0 2.5 3.0 3.5 4.0 4.5 5.0
consequently affects the definition of the peri-
Light intensity
meter of the particle projection. In this case, the
Figure 1. Influence of illumination intensity on use of threshold values below 55% don’t permit
shape factor estimates for spherical particles (A) and the correctly identifying of all pixels making up
irregular particles (B). the perimeter. This problem is noted with both

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006

 MICROSCOPIC IMAGE ANALYSIS TECHNIQUES 351

Vr rather more complex: 24 bits RGB (red-green-

A Vp blue) images and 32 bits CMYK (cyan-magenta-
er
yellow-black) images have formed by 3 and
AR
20 4 channels of 8 bits per channel allowing for a
256-level intensity scale for each channel. This
means that ‘‘color thresholding’’ requires estab-
15 lishment of a threshold value for each of the 3 or
Vr. Vp, er , AR

4 channels. In the context of particle size and shape

analysis, however, detailed color information is
10 less important than good definition of the particle
edge, allowing precise estimation of perimeter and
area. In fact, excellent results can be obtained by
5
ignoring the color information and considering
only pixel intensity.
0
40 50 60 70
Effects of Image Resolution (Pixel Size)
Threshold (%; 0 = white, 100 = black)
As noted by Zingerman et al. [1992] and Podczeck
et al. [1999],7,9 a key aspect of image analysis is
B 30 pixel size, governing the number of pixels making
Vr
Vp
up the particle projection. Pixel size may be varied
er either by modifying optical magnification or by
25
AR modifying digital image resolution. The results
presented so far (Figs. 1 and 2) were all obtained
20 with a JVC TK-S350 videocamera using the
Vr, Vp, e r , AR

minimum magnification (1
) permitted by the

15
10
5 reduced to 2.08 mm by using the highest optical
40 50 60 70
Threshold (%; 0 = white, 100 = black)
Figure 2. Influence of grey scale thresholding value
on shape factor estimates for spherical particles (A) and
irregular particles (B).
spherical and irregular particles, though more
frequently with the latter.
In the other hand the threshold selection
procedure will differ depending on whether we
start from a greyscale image or color image, and in
the latter case depending on whether we consider
color differences or merely differences in pixel
intensity. The thresholding procedure for an 8-bit-
per-pixel greyscale image, used in our study, is
straightforward, and simply involves assigning a
threshold value (on a greyscale of 0 to 255) above
which a pixel is considered as black. Nevertheless,
thresholding procedures for color images are
Olympus SZ-CTV stereomicroscope, allowing
obtention of greyscale images at 176
105 ppi
(Table 1). Under these conditions and using the
lowest optical magnification, each image repre-
sents a field of side 12.25 mm, which can be
magnification (6.3
). However, as noted by
Zingerman [1992],7 the use of high optical magni-
fications dramatically reduces depth of field,
hindering simultaneous focusing on all particles,
and often leading to poor edge definition and
over-estimation of particle dimensions. In addi-
tion, increased magnification also reduces the
field of view for this reason magnification is
generally used.
However, in recent years, there have been great
advances in digital photography, and the video
cameras and graphics cards previously used for
image capture have increasingly given way to
digital cameras. Images obtained with digital
cameras have high resolution and can be readily
transferred to the computer; in addition, they offer
a choice between various image qualities and file
types. It might be thought that under these
conditions the ideal would be to use the highest
resolution (smallest pixel size). However, higher-
resolution photographs clearly constitute much
larger files, so that the IA program needs to

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006

352 ALMEIDA-PRIETO, BLANCO-MÉNDEZ, AND OTERO-ESPINAR

Table 1. Basic Characteristics of the Images and Image Files Obtained by the Three Cameras Used in the Present
Study

Resolution PPI Real Scaling File

Image (Number (Pixels Image Pixel Size Factor File Compression Size
Camera Quality of Pixels) per Inch) Size (mm) (mm) (mm/pı́xel) Type Level (MB)
JVC – 704
420 176
105 101.6
101.6 144.3
241.9 12.25 TIFF None 0.29
TK-S350
Olympus SQ 640
512 72
72 225.8
180.6 352.8 25.64 JPG Low 0.15
DP11
SQ 1,280
1024 144
144 225.8
180.6 176.4 12.82 JPG Low 0.42
HQ 1,712
1368 144
144 301.9
241.3 176.4 9.69 JPG High 0.54
SHQ 1,712
1368 144
144 301.9
241.3 176.4 9.69 JPG Moderate 1.7
Low 6.9
Olympus 20.69a Low 0.20
DP12
SQ 640
480 72
72 225.7
169.3 352.8 39.98 JPG High 0.06
39.98 Low 0.20
13.89a Low 0.5
SQ 1,024
768 72
72 361.2
270.9 352.8 27.02 JPG High 0.15
27.02 Low 0.5
7.04a None 0.7
HQ 2,048
1536 72
72 722.5
541.9 352.8 13.33 JPG None 0.7
13.33 JPG Low 1.88
SHQ 2,048
1536 72
72 722.5
541.9 352.8 13.33 TIFF None 9.25
LQ, low quality; HQ, high quality; SHQ, super high quality.
a
Images obtained using the stereomicroscope at 1
magnification.

manage a large number of pixels, which greatly
increases the time required for the analysis. Thus
it is desirable to determine the minimum resolu-
tion at which particle dimensions can be accu-
rately measured.
To study and compare the effects of image
resolution and image file type on the different
parameters available for characterization of pellet
morphology, we obtained images of 50 pellets
using two models of Olympus digital camera, with
different image capture resolutions, different file
formats (TIFF or JPG), and different levels of JPG
data compression (see Table 1). In addition, we
obtained images with a wide-angle attachment
(0.5
) that increases the field of view and thus the
number of particles included in each image.
As can be seen from Table 1, with the three
systems four pixel sizes can be obtained, depend-
ing on the resolution of the system used. The DP12
digital camera has the lowest pixel density
(72
72 ppi) but the highest resolution (i.e.,
greatest total number of pixels (2,048
1,536)),
thus giving the smallest pixel size. The video
camera has a lower resolution (704
420 pixels)
but a high pixel density (176
105 ppi), and so
pixel size is lower than in most of the digital
camera resolutions. In any case, all of the resolu-
tions used (except for the lowest resolution of the
DP12 camera, 640*480 when used with the 0.5

attachment) gave pixel size after scaling lower
than the maximum of 30 mm recommended by
Podczeck et al. [1999]9 for the characterization of
pellets of around 1 mm diameter.
Figure 3 shows how number of pixels making up
the perimeter of the particle projection declines
exponentially with increasing pixel size. Bearing
in mind the results obtained by Almeida et al.
[2004],11 good precision requires that the radius
of the particle should correspond to at last 25–
30 pixels, or equivalently that the perimeter
should be made up of more than 150 pixels. This
suggests that we must reject resolutions with a
scaling factor of less than 20 mm per pixel.
To evaluate this possibility, we have plotted
various commonly used shape factors against
number of pixels making up the perimeter
(Fig. 4). As can be seen from these plots, all shape
factors except AR were affected to greater or lesser
extent by number of pixels making up the
perimeter, though particle shape (spherical or
irregular) was also important. With spherical
particles, the shape factors Vr and er showed a
rapid decline with increasing number of pixels
making up perimeter, until about 150–200 pixels,

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006

 MICROSCOPIC IMAGE ANALYSIS TECHNIQUES 353

550
after which the decline was less marked. Circu-
500 A, spherical particles larity likewise declined with increasing number of
B, irregular particles
pixels making up perimeter, while Vp increased.
Number of pixels in perimeter

450

400
An increase in the number of pixels making up
the perimeter (due to smaller pixel size) gives a
350
larger number of effective radii for calculation of
300
Vr, and allows better definition of the surface
250 roughness of the particles, thus causing a reduc-
200 tion in the values of those shape factors that are
150 sensitive to surface roughness (circularity and er),
100
and increased Vp. In view of the results obtained,
we can confirm that the use of scaling factors lower
50
5 10 15 20 25 30 35 40 than 20 mm/pixel should be sufficient for precise
Scaling factor (µm/píxel) estimation of most of the shape factors considered
here. This indicates that the DP11 and DP12 digi-
Figure 3. Number of pixels making up the perimeter tal cameras, with 1
optical magnification, can be
of the particle projection, plotted against the scaling used at any resolution except the lowest (640*512
factor used (mm/pixel). A, spherical particles; B, irregu-
and 640*480 in DP11 and DP12 respectively),
lar particles.

Spherical particles 13 0.92

Spherical particles Spherical particles
Irregular particles Irregular particles
16 Irregular particles
12 0.90

11 0.88
15
Circularity

10
0.86
Vp
Vr

0.80
7
5
0.78

0.76
6
4
0.74

100 200 300 400 500 100 200 300 400 500 100 200 300 400 500
No. of pixels in perimeter No. of pixels in perimeter No. of pixels in perimeter

Spherical particles 0.58

1.64
Spherical particles
Irregular particles Irregular particles
1.62
0.56
1.60

1.58 0.54

1.56
0.52
AR

er

0.25
1.11

0.20

1.10
100 200 300 400 500 100 200 300 400 500
No. of pixels in perimeter No. of pixels in perimeter

Figure 4. Influence of number of pixels making up the perimeter of the particle

projection on estimates of the various shape factors.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006

354 ALMEIDA-PRIETO, BLANCO-MÉNDEZ, AND OTERO-ESPINAR
although use of the 0.5
attachment on DP12
would require the highest resolution (2048*1536).
File Format and Compression Ratio
Another important aspect to be taken into account
is file format. In the case of the system based on
the video camera, this aspect is not so critical,
since it is not necessary to save the images before
analysis; rather, the signal can simply be frozen
and the particles measured directly from the
frozen image. With the digital cameras, however,
it is necessary to store the images Under these
conditions image size is important, since it affects
download time, storage space requirement and
subsequent image treatment.
To empirically investigate the effects of file
format on shape factor values, we obtained several
identical photographs of a single field of view
containing 20 particles, using different resolutions
and file formats. In each case sextuplicate images
were obtained. This part of the study was carried
out using the Olympus DP12 camera, which allows
various resolutions, storage in TIFF or JPG
format, and various compression ratios in JPG
format (see Table 1). TIFF (Tagged Image File
Format ) is a high-quality uncompressed format
offering up to 64-bit color, and thus basically
complete image information; the disadvantage is
that the files may be very large (up to 4 GB). JPG or
JPEG (Joint Photographic Experts Group) takes
into account the imperfections of the human visual
system as regards sensing of certain colors and
other information, and eliminates this information
to reduce file size. Notably, the JPG compression
algorithm eliminates minor color differences
between adjacent pixels; at low compression ratios
these changes are not visually detectable, though
Table 2. Mean Radius and Shape Factor Values Obtained in Analysis of Particle Batches Using Different File
Formats (TIFF or JPEG) and Different JPEG Compression Ratios (Low or High)
1024
768
JPEG
Low
Particles measured 120
Number of pixels in perimeter 115.13  0.23
Mean radius (mm) 547.79  0.18
er 0.57  0.005
Vr 4.56  0.06
Vp 6.31  0.23
Values shown are means  standard deviations for 72–120 particles.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006
they may become apparent at higher compression
ratios. In any case, color information is not
necessary for characterization of particle shape,
so that a priori the JPG compression algorithm
would not be expected to affect shape factor
estimates, except at high compression ratios. The
results obtained are summarized in Table 2.
Analysis of variance indicated significant differ-
ences in Vr (a < 0.01) between the 2048*1536 and
1024*768 resolutions, but no significant effects of
file format or JPG compression ratio. In the case of
Vp, er, and circularity, however, not only resolu-
tion but also file format and JPG compression ratio
had significant effects. Finally, AR was not
significantly affected by any of these factors.
To identify the causes of these effects, we
performed a similar study of the parameters
directly measured by the IA program, namely
area, perimeter and mean radius. Analysis of
variance indicated that both area and mean radius
are basically independent of resolution, file format
and compression ratio; however, perimeter was
significantly affected (a < 0.05) by both file format
and JPG compression ratio. This explains why Vp,
er, and circularity (which are based on the
perimeter measurement) are affected by these
factors. The effects of file format and JPG compres-
sion ratio on perimeter are related to the way in
which the IA program identifies the pixels that
make up the perimeter. Under the illumination
and thresholding conditions used here, an exces-
sive increase in resolution or in amount of
information stored greatly hinders the process of
identification of the perimetral pixels, because an
increase in image information (i.e., more pixels
with more color and/or intensity information)
means that the transition between particle projec-
tion and background is more gradual. This means
Resolution-File-Compression
1024
768 2048
1536 2048
1536
JPEG JPEG TIFF
High — —
120 97 72
114.99  0.13 233.80  3.80 237.99  3.87
548.35  0.17 541.12  7.85 546.62  7.14
0.56  0.006 0.53  0.011 0.55  0.018
4.51  0.04 4.23  0.32 4.13  0.37
5.84  0.01 7.96  0.49 8.75  1.15
 MICROSCOPIC IMAGE ANALYSIS TECHNIQUES 355

in turn that, as can be seen from Table 2, the
number of particles for which the perimeter can be
completely identified declines markedly with
increasing image quality. To attempt to resolve
this problem it would be necessary to re-optimize
illumination and thresholding for high-resolution
images.
In fact, the effects of file format and JPG
compression ratio, although statistically signifi-
cant for some shape factors, were of scant practical
relevance: the modifications were in all cases of
less than 5% (maximum 1.3%, 4.0%, 0.4%, and
2.9% for mean radius, er, Vr, and Vp, respectively).
As noted, however, the number of particles for
which the perimeter is correctly identified dropped
markedly at higher resolutions, making it neces-
sary to have a larger total sample size, and
requiring operator intervention to ensure that
only particles with correct perimeter identification
are being included: thus we consider that it is
better to use lower-resolution images.
Color Versus Black-and-White Images
Another relevant aspect is whether parameter
estimation is significantly affected by color
information, that is 24-bit-per-pixel RGB format
or 8-bit-per-pixel greyscale. A priori, and bearing
in mind that conversion to greyscale maintains all
color and intensity information, with only color
information lost, we would expect only slight
effects.
To empirically investigate the magnitude of
these effects, we again used the Olympus DP12
camera, at three resolutions (640*480, 1024*768,
and 2048*1536), obtaining photographs of 50
spherical particles and 50 irregular particles in
both color and greyscale. The results obtained
(Table 3) confirm that loss of color information has
minimal impact of shape factor values.
Repeatability and Precision
Having identified the illumination and image-
processing factors with most marked effects on
shape factor values, we next proceeded to evalu-
ate method repeatability and precision. First, we
evaluated variability among particle batches
characterized on different days. To this end, we
randomly selected 10 particles from each batch of
particles used in the study. Images were obtained
daily for 3 days with a video camera (in this

Table 3. Influence of Color Information (Color or Black-and-white) and Image Resolution on Estimates of Mean
Radius and the Various Shape Factors

Spherical Particles

Color B and W
Resolution Resolution

640
480 1,024
768 2,048
1536 640
480 1,024
768 2,048
1,536
Number of 77.55  0.33 115.44  0.06 225.36  3.05 75.10  0.14 112.72  0.14 224.99  1.65
pixels in
perimeter
Mean radius 546.39  0.34 549.29  0.56 531.24  0.75 530.81  0.24 536.00  0.11 525.34  1.78
(mm)
Vr 5.71  0.05 4.75  0.06 4.32  0.08 5.52  0.05 4.95  0.05 4.34  0.01
Vp 6.10  0.17 5.98  0.08 7.09  0.22 5.52  0.29 6.05  0.10 6.55  0.11
er 0.57  0.004 0.55  0.001 0.55  0.002 0.57  0.005 0.56  0.001 0.55  0.001
Irregular particles
Number of 90.01  0.32 134.07  0.16 266.53  2.91 89.65  0.58 133.93  0.24 270.09  1.34
pixels in
perimeter
Mean radius 607.10  0.04 611.41  0.31 595.38  2.83 604.87  1.69 609.42  0.27 601.06  2.39
(mm)
Vr 15.08  0.004 14.89  0.001 15.34  0.22 15.23  0.13 14.99  0.01 14.97  0.24
Vp 9.99  0.007 10.62  0.14 11.33  0.21 10.09  0.06 10.59  0.20 11.45  0.28
er 0.28  0.001 0.25  0.001 0.24  0.009 0.26  0.008 0.25  0.002 0.24  0.008
Values shown are means  standard deviations.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006

356 ALMEIDA-PRIETO, BLANCO-MÉNDEZ, AND OTERO-ESPINAR
case JV TK-S3501), then digitalized and analyzed
to obtain particle measurements. The results
obtained indicate that particle measurements
varied by at most 5–8% between days, while
repeated measurements of a single day’s images
did not give rise to any variation. This suggests
that the observed variation among days is
probably attributable to slight differences in
particle positions in the field of view. Variations
in particle position can be expected to affect the
angle of incidence of the illumination, thus
affecting brightness and contrast of the resulting
images, because cameras of this type determine
brightness and contrast automatically, depending
on the illumination received by the particles
(which depends on their precise position). How-
ever, this type of error can be minimized by
increasing the number of particles measured,
since their position in the field of view is random.
That the magnitude of the error depends on
particle position and thus light reflected is
confirmed by the fact that error is minimal if
repeated images are obtained consecutively.
Sample Size
Finally, to assess the best mode of operation
for morphological particle analysis it is necessary
to determine the minimum number of particles
that need to sampled in order to be accurately
representative of the population. This requires
estimation of population variability on the basis
of sample variance, and pre-establishment of
the minimum detectable between-population dif-
ference for a given significance level and statis-
tical power. To this end we used the next
equation:12
j k
n  ðza þ zb Þ2 s2 =d2
where n ¼ number of particles to be sampled,
s ¼ sample variance, and d ¼ the minimum differ-
ence to be detected. For estimation of sample
variance, we selected the particle batch showing
highest dispersion of Vr and Vp. Using this
estimate, and for a significance level of 0.05 (a)
and a power of 0.05 (b), we found the minimum
sample size requirement to be 300 particles,
allowing detection of differences of as little as
1.25 and 0.73 for Vr and Vp in respectively, clearly
sufficient for accurate characterization of particle
morphology. However a minor sample could be
required in the case of more homogeneous
populations.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006
CONCLUSIONS
As other authors have commented previously,
estimates of various shape factors used for the
morphological characterization of particles may
be markedly influenced by diverse process vari-
ables related to illumination (position of the light
source, or light intensity), the greyscale threshold
value used for distinguishing between particle
and background, and pixel size. As a result, before
attempting any such morphological characteriza-
tion it is essential to identify and standardize
optimal conditions for the equipment being used
and for the type of particle to be characterized. In
addition, the use of digital cameras and image
analysis software requires decisions about resolu-
tion, file format and compression ratios, which
may again affect shape factor estimates. The
results of the present study indicate that the best
results are not necessarily obtained using the
highest image resolution available. With our
photographic system we have found that the use
of moderate resolutions and greyscale images
facilitates many steps in image analysis, includ-
ing thresholding and identification of the pixels
making up the perimeter of the particle projec-
tion.
REFERENCES
1. Cox EP. 1927. A method of assigning numerical and
percentage values to the degree of roundness of
sand grains. J Paleont 1:179–183.
2. Schneiderhöhn P. 1954. Eine verleichende studie
über methoden zur quantitativen bestimmung von
abrundung und form an sandkörnern. Heidelber-
ger Beiträge zur Mineralogie und Petrographie
4:172–191.
3. Hausner HH. 1966. Characterization of the powder
particle shape. Planseeberichte für pulvermetal-
lurgie 14:75–84.
4. Exner HE, Link E. 1977. Image analysis for
characterization of size and shape distribution of
load powders. Pow Met Int 9:131–133.
5. Barret PJ. 1980. The shape of rock particles, a
critical review. Sedimentology 27:291–303.
6. Lövgren K, Lundberg PJ. 1989. Determination of
sphericity of pellets prepared by extrusion/spher-
onization and the impact of some process para-
meters. Drug Dev Ind Pharm 15:2375–2392.
7. Zingerman JP, Mehta SC, Salter JM, Radebaugh
GH. 1992. Validation of a computerized image
analysis system for particle size determination.
Pharmaceutical applications. Int J Pharm 88:
303–312.

8. Podczeck F, Newton MJ. 1994. A shape factor to
characterize the quality of spheroids. J Pharm
Pharmacol 46:82–85.
9. Podczeck F, Rahman SH, Newton MJ. 1999.
Evaluation of a standardised procedure to assess
the shape of pellets using image analysis. Int J
Pharm 19:123–138.
10. Podczeck F, Newton MJ. 1995. The evaluation of a
three-dimensional shape factor for the quantitative
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 2, FEBRUARY 2006
MICROSCOPIC IMAGE ANALYSIS TECHNIQUES 357
assessment of the sphericity and surface roughness
of pellets. Int J Pharm 124:253–259.
11. Almeida-Prieto S, Blanco-Méndez J, Otero-Espinar
FJ. 2004. Image analysis of the shape of granulated
powder grains. J Pharm Sci 93:621–634.
12. Walpole RE, Myers RH, Myers SL. 1998. One and
two sample test of hypotheses. In: Heath A, editors.
Probability and stastatistics. 6th Ed., New Jersey:
Prentice Hall Inc. pp 290–357.