Digital dental photography.

IN BRIEF
We live in a digital world, and recent

Part 3: principles of technological advances have offered

PRACTICE
conveniences and facilities that were
once only stuff of dreams

digital photography The eyes and digital sensors share
uncanny similarities, unlike film
photography that is rigid and inflexible
Digital photography can be summarised
1 by the acronym CPD (capture, processing
I. Ahmad and display).

VERIFIABLE CPD PAPER

Although we live in a digital age, our knowledge of the processes and technology involved is often limited. As a foundation
to understanding the subsequent parts of this series, this part describes the fundamental aspects of digital photography,
which includes the sensors, processing and display.

We are currently living in the digital revo- passes continuously, without divisions an electrical current, or charge, which is
lution: digital broadcasting, digital con- or separations. eventually processed into an image. Ocular
sumer goods, digital dental radiography, In a similar vein, digital photography and digital imagery share many similari-
and photography is no exception. Without offers many benefits compared to conven- ties and are both extremely flexible. For
doubt, digital is the future. tional photography including: example, if we see something we do not
However, the natural world is analogue; Instantaneity and convenience like, we can look away (with digital pho-
everything around us is continuous: col- Flexibility for editing, copying and tography, unwanted parts of an image
our, space, time and sound are all sinu- disseminating images can be cropped). If something attracts our
ous, without discrete separations. We have Environmentally greener by attention, the eyes concentrate on that
separated nature, or digitised it, for the eliminating toxic dyes and processing specific part of the object or subject (with
purpose of convenience, utilisation and chemicals digital photography, any point of interest
manipulation. An example is time, which Long-term economy by reusing storage can be enlarged). Also, if we do not like
we have divided into days, hours, minutes media such as memory cards. what we see, the brain can change the con-
and seconds, but which in reality, similarly text of reality so that we find the apparent
to our surroundings, is not intermittent but As with any new technology, there is, unsightly representation more pleasurable
however, a learning curve to fully utilise (with digital photography, software manip-
the benefits and avoid the pitfalls. This ulation can alter an image to any desired
FUNDAMENTALS OF DIGITAL chapter describes the fundamental aspects parameter). These few examples highlight
DENTAL PHOTOGRAPHY of digital photography, which serve as an the flexibility and uncanny similarities of
essential foundation for subsequent chap- ocular and digital imagery.
1. Digital dental photography: an overview
ters. The starting point is the quintessential Conversely, chemical or film photogra-
2. Purposes and uses
item for digital photography: the sensors. phy is rigid, with little scope for manipula-
3. Principles of digital photography
tion and therefore requires that all settings
4. Choosing a camera and accessories THE SENSORS be exact if an acceptable image is to be
5. Lighting
Light sensors can be categorised into three produced. The basis of chemical photog-
6. Camera settings
basic types, ocular, digital and chemical. raphy is photosensitive coloured layers
7. Extra-oral set-ups
Surprisingly, the fundamental principles of painted onto a film emulsion which, fol-
8. Intra-oral set-ups
the three types are very similar. The ocu- lowing development, reveals the registered
9. Post-image capture processing
lar apparatus consists of the eyes, optic image on a cellulose sheet. To produce a
10. Printing, publishing and presentations
nerve and the brain, which is the ultimate correctly exposed and high quality image,
arbitrator for assessment irrespective of the every setting needs to be accurate. For
1
General Dental Practitioner, The Ridgeway Dental
method used to create an image. In the example, sharp focusing, correct orienta-
Surgery, 173 The Ridgeway, North Harrow, light sensitive retina of the eyes, coloured tion, proper framing and composition, pre-
Middlesex, HA2 7DF
Correspondence to: Irfan Ahmad
dyes are stimulated by incoming light, cise aperture opening and shutter speeds.
Email: iahmadbds@aol.com triggering neural responses to the brain, In addition, the colour temperature of the
www.IrfanAhmadTRDS.co.uk
which subsequently computes the image ambient light must match that of the film
Refereed Paper of the object being viewed. emulsion, and the developing chemicals
Accepted 15 November 2008
DOI: 10.1038/sj.bdj.2009.416
In digital photography, light sensitive need to be precisely diluted and at the
British Dental Journal 2009; 206: 517-523 diodes act as the sensors, which create correct working temperature. It is obvious

BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009 517

2009 Macmillan Publishers Limited. All rights reserved.
PRACTICE

that with so many variables, the scope

of error is magnified and even if camera Table 1 Comparison of the three types of sensors
settings are correct, incorrect developing
Ocular Digital Chemical
can produce unsatisfactory results. Besides
the obvious convenience and instance of Light sensitive sensor Retina Electrical diodes (pixels) Dyes on film sheets
digital photography, a major advantage is Storage/relay media Nerves Memory cards or disc Film sheet
the ability to correct technical errors at a
Reproduction media Brain Monitor, projector, print Film sheet
later stage using software manipulation for
rectifying exposure, white balance, fram-
ing, orientation, sharpening, etc.
A comparison of the three sensors, ocu-
lar, digital and chemical, is summarised in
Table 1. It is worth noting that for chemical
photography, the film sheet serves as the
light sensitive sensor, storage and repro-
duction media. However, with the ocular Pixel
and digital imagery, each of these three
entities uses different media, which obvi-
ously expands possibilities for manipula-
tion and offers unparalleled flexibility.

TECHNICAL ASPECTS Fig. 1 Full-frame CCD with a large fill factor

OF DIGITAL PHOTOGRAPHY
The easiest way to describe the basic prin-
ciples of digital photography is by divid- Conducting
ing them into three processes, forming the area
acronym CPD:
C for capture
P for processing
D for display. Pixel

Capture
The heart of all products based on sili-
cone technology, such as computers, stor-
age media, scanners and digital cameras, is
a semiconductor. With image sensors, the Fig. 2 Interline CCD with a small fill factor
semiconductors are photosensitive units
composed of tiny light detecting units Micro-lens
called pixels. The latter are a substitute
for emulsion in conventional film cameras.
Conducting
Pixels come in many shapes and qualities,
area
varying in size from 5 m to 12 m.
Basically, the image sensor is a col-
lection of silicone photodiodes (pixels),
which register the intensity of brightness
Pixel
and darkness of an object. In effect, they
are only capable of producing a black and
white image of the object being photo-
graphed. To create a colour image requires
using appropriate filters corresponding to
the three additive primary colours red,
green and blue. Fig. 3 Interline CCD with micro-lenses, which increase the fill factor
Currently there are two types of image
sensors competing in the market, the CCD The CCDs can be further divided into full- captured onto the sensor. The pixels are
(charged coupled device) and CMOS (com- frame and interline CCDs (Figs 1-2). The arranged in a line, and once stimulated
plementary metal oxide semiconductors), former, full-frame CCDs, allow the entire by light, convey the electrical charge to
each having advantages and disadvantages. frame viewed in the viewfinder to be the end of the line where it is processed

518 BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009

2009 Macmillan Publishers Limited. All rights reserved.
 PRACTICE

to form an image. This procedure is time

consuming and must be carried out in
darkness, that is, after the camera shut-
ter is closed. To expedite the process, the
interline CCDs have non-light sensitive
rows between the pixels which convey
the electrical charge simultaneously as
the pixels are stimulated by the incoming Fig. 4 A small bit depth results in a pronounced jagged edge
light. This accelerates the process of creat-
ing an image, but due to the conducting
row (non-light sensitive areas), the light
sensitive area (fill factor) available is lower
compared to the full-frame CCDs. This is a
major advantage of full-frame CCDs, since
a fill factor of 70% to 90% means that less
image information is lost compared to an
interline CCD, which has a fill factor of Fig. 5 Increasing the bit depth creates a smoother edge
30% to 50%.
The other competitors to CCDs are the
CMOS sensors. These devices register light
similarly to a CCD, but processing is per-
formed on each pixel rather than being
conducted to the end of a line. Due to
the circuit integration within each pixel,
the CMOS sensors have a smaller surface
area with a reduced fill factor of only Fig. 6 To create a seamless transition between black and white and a smooth edge, a minimum
bit depth of 8, or 256 levels, is necessary
30%. Other limitations of CMOS sensors
are a low dynamic range and increased
noise levels, which are both detrimen- software (for example, Adobe Photoshop). for red, 8 bit for green and 8 bit for blue
tal to image quality. The advantages are Since all digital equipment and software (256 levels for red, 256 levels for green,
lower power consumption and elimination work with a binary code, a transgression and 256 levels for blue), or a total of 24
of booming (overflow of excess electrical about digitisation is essential. bits (224). This is referred to as the bit or
charges to adjacent pixels). To circumvent All computers function with a digital colour depth of an image. The greater the
the low fill factor of both interline CCDs binary code, that is, they can only com- bit depth per primary colour, the greater
and CMOS sensors, either micro-lenses or prehend 0 and 1. These binary digits are the accuracy of recorded detail. The colour
octagonal shaped pixels with larger sur- termed bits, and eight bits are equivalent depth is an important point to consider
face areas arranged diagonally are used to to 1 byte. A byte is the minimum number when purchasing digital equipment or soft-
increase the light-sensitive potential or fill of bits required to make up a single alpha- ware. The stated colour depth can either be
factor (Fig. 3). betic character. The storage and memory for each primary colour (per channel), or
capacities of computers are therefore the total bit depth of the three primary
Processing quoted in megabytes (MB) or gigabytes colours. Many manufacturers quote 8-bit
There is a misconception equating an image (GB). In digital photography the primary depth/channel, indicating 8 bit per primary
sensor with digital images. However, the function of the A/D converter is to repro- colour (that is, 8 for red, 8 for green and
contrary is true. A sensor is only capable of duce the pure analogue signal into a dig- 8 for blue), which equates to a total bit
delivering an analogue signal. This signal ital code that is as close as possible to the depth of 24. However, other manufacturers
is an electrical charge, obtained from the original. Since digital data is composed state the total bit depth of say, 24.
result of exposing the pixels to light. Each of discrete entities, lacking homogeneity Returning to the image sensor, each
pixel creates an electrical charge depending with noticeable banding, in order for the pixel is assigned a binary number accord-
on the intensity of incoming light and the human eye to visualise a continuous tonal ing to the magnitude of its charge. The A/D
duration of exposure. Further technology is range (greyscale), a minimum of 8 bits is converter assigns the level of brightness in
necessary to transform the analogue elec- necessary (28), which equate to 256 levels steps. As mentioned above, the greater the
trical signal into binary (digital) numbers. (Figs 4-6). number of steps the smoother the transi-
This is achieved by an analogue-digital Consequently, for colour images com- tion on a greyscale and the more precise
converter (A/D converter). Once converted, prising the three primary colours red, green rendition of an image. Most cameras use
the digital data is processed by microchips and blue, each colour channel must have an 8 bit per channel A/D converter cod-
either within the camera, or downloaded a minimum of 8 bits so that a continuous ing for 256 different levels of brightness
to a PC and manipulated with appropriate tonal range is perceived. This means 8 bit and darkness, while professional systems

BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009 519

2009 Macmillan Publishers Limited. All rights reserved.
PRACTICE

use 16 bit per channel, which translates

to 65,536 brightness levels. To calculate
the number of colours possible in a given
system, the tonal levels for each colour
are multiplied. For example, for a cam-
era with a total colour depth of 24, the
number of colours coded is 256R 256G
256B, which results in 16.7 million pos-
8 bit (28 = 256)
sible colours (Fig. 7), and for a total 48-bit
system (16 bit/channel) or 248, the number
of colours is 2.8014. In comparison, the dif-
ference threshold for colour of the human
eye is low enough to discriminate 7 mil-
lion colours.
In reality, only 8 bits are necessary for
the eyes to visualise an uninterrupted
smooth greyscale. However, once an 8-bit
image is manipulated using photo editing
Red channel (28 = 256) Green channel (28 = 256) Blue channel (28 = 256)
software, there is degradation of the origi-
nal 8-bit signal resulting in jagged steps
at the periphery of objects. To avoid these
unwanted artefacts, it is therefore wiser
to start with a 16 bit image, allowing for
degradation while still maintaining the
minimum requisite 8 bit colour depth.
Besides colour depth, the other factor
to consider is the dynamic range of the
sensor (to be discussed further in Part 6).
16.7 million colours
This is determined by the amount of charge
that a pixel can accept, or its saturation Fig. 7 Schematic representation of a 24 bit colour depth system (or 8 bit/channel)
level, termed full well capacity. The larger
Bayer pattern
the physical size of a pixel, the greater
the charge it can hold, and the greater the
dynamic range. Therefore, in high-end
digital cameras the sensors have larger
pixels of approximately 12 m, compared
to the 5 m pixels that are used in amateur
or compact cameras.
As previously stated, the image captured
by a sensor is in essence black-and-white
(Fig. 8). Colour is achieved by adding three
channels representing the three primary
colours of additive mixing, that is red,
green and blue. A variety of ingenious
methods are utilised for creating coloured
images including rotating colour filters,
beam splitters, or coating each pixel with
filters of the three primary colours. The lat-
ter is the most popular method, creating a Fig. 8 The pixels of an image sensor only Fig. 9 Colour is captured by placing filters of
mosaic of red, green and blue on the image detect the brightness levels of an object, red, green and blue onto the pixels, such as
creating a black and white image this Bayer pattern arrangement
sensor (Fig. 9). When exposed, each pixel
registers the intensity of light for one of the
three colours, and when collated together The rationale is that the red, green and white light, each individual layer registers
this yields a colour image (Fig. 10). The blue components of white light penetrate red, green or blue, and when combined
latest technology is sandwiching three at different depths: red is deepest, green they form a coloured image.
separate pixel layers of red, green, and intermediate and blue superficial. When The actual image capture is performed by
blue, similar to the dye emulsions of film. these multi-layer sensors are exposed to numerous methods, for example scanning,

520 BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009

2009 Macmillan Publishers Limited. All rights reserved.
 PRACTICE

Missing inter-pixel information

is interpolated from adjacent pixels

11 12

Fig. 10 The filtered pixels record the amount

of red, green and blue at each site to form a
colour image

3-shot, 1-shot, 4-shot, microscan and

macroscan. The method used depends on
the application in question. For example,
for photographing static still-life compo-
sitions or documenting works of art and
sculptures, the ideal is the scanner system.
However, a scanning system is inappro-
priate for moving objects, and for sports
photography the 1-shot system is the
13 14
ideal choice.
By far the most popular system, and one Figs 11-15 The pixels are only capable of
that is suitable for dental photography, is measuring the dark and bright parts of an
the 1-shot system. As the name implies, a image, and in effect are only capable of
producing a black and white image (Fig. 11).
single exposure is required to capture the
Colour is created by the three channels red
object being photographed. The set-up is (Fig. 12), green (Fig. 13) and blue (Fig. 14),
as follows. The image sensor consists of and combining these channels produces a
pixels with a mosaic of filters, for example colour image (Fig. 15)
in the Bayer pattern arrangement for the
three primary colours, red, green and blue
(Fig. 9). Once exposed, the sensor records
the corresponding amount of red, green
and blue in the prevailing composition.
The entire process is summarised in Figs
11-15. Depending on the proximity of the
pixels, small amounts of detail are lost,
which are interpolated using information
from adjacent pixels. The disadvantages of 15
this system are that the resulting image is
prone to interpolation errors such as col-
our fringes at edges of objects and Moir errors by using software to suppress these while a lengthy computation time may be
patterns of strongly chequered materials. unwanted artefacts. However, if suppres- unacceptable to expedite workflow. But
The interpolation algorithms mitigate these sion is too great, genuine detail is lost, taking all factors into consideration, the

BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009 521

2009 Macmillan Publishers Limited. All rights reserved.
PRACTICE

Fig. 17 An image with a file size of 30.3 MB

and 2,797 1,895 pixels (5.3 megapixels)

Fig. 16 LCD display on camera back

benefits of a 1-shot system, such as the loss, but is easier to store, manipulate and Fig. 18 An image with a file size of
113.4 MB and 5,329 3,717 pixels
ability to record moving subjects, com- disseminate. The JPEG format also has a (19.8 megapixels)
pactness, lightweight and reduced cost, range of resolutions from low to high, with
outweigh the minor and perhaps imper- corresponding file sizes, respectively.
ceptible loss in image quality. If a proprietary format is chosen, the file
Before an image can be viewed a certain is in a raw state and requires processing by
amount of processing is necessary. Firstly, specific software before it can be viewed
the captured image must be processed by and stored into a generic format. On the
software in the camera as a digital file. The other hand if a generic format is chosen at
format of the file at this stage can either be the outset, no further processing is neces-
proprietary, that is specific to a particular sary to view or store the file.
camera manufacturer, or in a generic format
such as RAW (PNG), TIFF or JPEG. Secondly, Display
the size of the ensuing file depends on After in-camera processing, the image
the format in which it is saved. The file can be displayed via electronic or printed
size is a crucial determinant of the final media. Electronic media consists of moni- Fig. 19 A 100% enlargement of a section
of the image shown in Figure 17, with a file
image quality. The file size of an image can tors and projectors, and printed media of size of 2.3 MB and 639 616 pixels
be calculated according to the formula: photographic paper or printing paper. (0.4 megapixels)
Number of pixels (total bit depth The first time that an image is usually
8) = image size in bytes viewed is on the LCD monitor on the cam-
era back (Fig. 16) The size of these moni-
For example, the maximum file size that tors varies from 2 inches to 3.5 inches,
a digital camera with a 10 million pixel with a resolution ranging from million
image sensor, and a bit depth of 24 (8 bits to 1 million pixels. The monitors allow
per primary colour) is capable of creating instantaneous viewing of the image for
in an uncompressed state is 30 MB: assessing composition, framing, orienta-
[10 106 (24 8)] = 30 MB tion and exposure. However, they are of
little use for determining fine detail or
The file format, and hence its size, is pri- sharp focus due to their low resolution,
marily dependant on the intended use of usually no more than 1 megapixels.
the image. This is usually in a RAW or TIFF The second type of electronic viewing
format. If a smaller size file is required, a is with a computer monitor or an LCD Fig. 20 A 100% enlargement of a section
low-resolution file format such as JPEG can projector (beamer). The resolution of both of the image shown in Figure 18, with a
be chosen. The latter format compresses the computer monitors and projectors varies file size of 7.6 MB and 1,140 1,153 pixels
(1.3 megapixels)
original digital file at the expense of detail enormously. For the former the resolution

522 BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009

2009 Macmillan Publishers Limited. All rights reserved.
 PRACTICE

ranges from as little as 720 480 (0.3 meg- The final point worth mentioning about the monitor (1.3 megapixels). However, if
apixels) to 1,440 900 (1.3 megapixels). megapixels and monitors or projectors is both images are now enlarged by 100%
Even the state-of-the-art true high defini- as follows. While both have relatively low to concentrate on the lower mandibular
tion projectors are only capable of deliver- pixel counts compared to digital cameras, incisors, the pixel count of the enlarged
ing a resolution of 2 megapixels (1,920 a difference is noticeable when an image is section of Figure 17 is 639 616 (0.4
1,080), far short of what is achievable with enlarged. For example, this is particularly megapixels) Figure 19, while for Figure
even the most inexpensive digital compact relevant when photographing pathologi- 18 it is 1,140 1,153 (1.3 megapixels)
cameras. This is the reason that an image cal changes to the oral mucosa. If a small Figure 20. With this enlargement, the pixel
taken with a 3 megapixels camera will lesion, in its early stages, is detected it is count of Figure 19 is lower (0.4 megapix-
look the same as that from a camera with useful to magnify the area for detailed els) than the monitor (1.3 megapixels),
a 10 megapixel sensor. If no difference visual assessment. However, if the image and the image appears grainy and is seen
is visually discernable, why bother with deteriorates when enlarged, it is clinically to be breaking down, with loss of detail.
expensive, high megapixels cameras? The useless, giving few clues to the pathologi- However, the enlarged image in Figure 20
reason is as follows: the resultant image cal process. To illustrate this point, con- matches the pixel count of the monitor
quality is not solely dependant on the sider the two images in Figures 17 and (1.3 megapixels) and still appears sharp
number of pixels. Other more important 18, which were taken with identical light- and retains detail. Notice the scratches
factors include resolving power of the lens, ing, lens, etc. but with digital backs of on the enamel surface on the mandibular
tonal range of the entire system such as bit different pixel count image sensors. The right lateral incisor in Figure 20, which are
depth, dynamic range, file format and size, image in Figure 17 is a 30.3 MB file, with indiscernible in Figure 19. This example
camera hardware (A/D converter, cooling), a pixel count of 2,797 1,895 (5.3 meg- emphasises the need to use camera equip-
and image processing software (interpola- apixels), while the image in Figure 18 is ment with high specifications, including a
tion and colour reproduction), etc. Hence a 113.4 MB file with a pixel count of high pixel count, to retain quality when an
an elaborate camera system usually offers 5,329 3,717 (19.8 megapixels), that is, image is enlarged.
more than just higher megapixels, but also the second image has a nearly four times The final method of viewing an image is
the features cited above to produce high greater pixel count. When both images are printing, which can either be with an office
quality images. This is an important point viewed full-frame on a standard, 1.3 meg- printer or a professional printing press.
to remember before choosing a camera sys- apixel computer monitor, no difference is Both methods are ubiquitously popular,
tem (covered in Part 4), because two cam- visually perceptible. The reason is that the each having their unique benefits and
eras with identical megapixels will produce pixel count of both images (5.3 megapix- drawbacks. Printing is discussed further
drastically different quality images. els and 19.8 megapixels) exceeds that of in Part 10.

BRITISH DENTAL JOURNAL VOLUME 206 NO. 10 MAY 23 2009 523

2009 Macmillan Publishers Limited. All rights reserved.