APPLICABILITY OF MOBILE PHONES FOR

TELE-DERMATOLOGY

A Pilot Study

Scheibböck Christian

, Dreiseitl Stephan

Vienna University of Technology, Karlsplatz 13, 1040 Vienna, Austria

Department of Software Engineering, Upper Austria University of Applied Sciences at Hagenberg

Softwarepark 11, 4232 Hagenberg, Austria

Binder Michael

Department of Dermatology, Division of General Dermatology, Medical University of Vienna, Austria

Keywords: Telemedicine, Tele-dermatology, Mobile phones, Camera phones, Colour fidelity, MMS, Illumination.

Abstract: Examination in dermatology is primarily based on visual inspection. Since this visual information can be

stored and transmitted easily through digital images, tele-consultation and tele-diagnosis are predestined

methods especially in the field of dermatology. Nowadays mobile phones represent the most important

communication tool around the world. To study how the process of acquiring, transmitting and diagnosing

images can be implemented by using cheap and widely available devices, such as mobile phones, we

conducted several experiments at the Medical University of Vienna. Material and Methods: In a study

patients were asked to take one or more photos of their dermatoses with the camera of a mobile phone.

Images were transmitted electronically by using the MMS function of the phones. All participants were

examined routinely at the outpatient department, establishing the gold standard diagnosis. Three different

phone models were evaluated with regard to colour fidelity, illumination effects, and terms of resolution.

Images for tests were produced at not standardized conditions. Results:

In all three phones high resolution

images are compressed. Resolution was skilled down to 640x480 while sending via MMS. Colour fidelity

was different depending on the manufacturer. Colour fidelity increased proportional with increasing

illumination.

1 INTRODUCTION

The use of telemedicine, especially tele-

dermatology, can be of great advantage for patients

and doctors. The following points show where tele-

dermatology can be of use:

 Outlying territories with a low density of

dermatologists

 Patients without access to a car or public

transportation

 Patients who want to save time

 Patients who are in foreign countries and want

to receive a diagnosis and therapy advice in

their native language

 Patients who want to avoid or shy personal

contact to doctors due to several reasons (e.g.

religion, agitation, shame, etc.)

 Patients who want to obtain a fast (and

reliable) second opinion

Some examples where tele-dermatology can be

of use for doctors:

 Second opinion: with telemedicine it is

possible to receive another specialized opinion

without time delay

 Temporally optimized advanced training:

images can be stored and used as instruction

material for medical students and doctors

Further possibilities of tele-dermatology:

 Selection of patients - the urgency of cases

can be judged, in order to achieve a faster and

more efficient treatment (triage)

474

Christian S., Stephan D. and Michael B. (2009).

APPLICABILITY OF MOBILE PHONES FOR TELE-DERMATOLOGY - A Pilot Study.

In Proceedings of the International Conference on Health Informatics, pages 474-477

DOI: 10.5220/0001550504740477

 SciTePress

 Optimal resource use through purposeful

transfer: patients mustn’t wait for

investigation at a “mismatching” department.

They are passed on to the appropriate

specialists.

 Follow-up assistance: the patient can receive

further assistance economically and fast (e.g.

wound inspection)

The university clinic has the advantage of always

being on the newest state of the art and science. This

results in some disadvantages for the patient, e.g.

long waiting times for an appointment at the

outpatient’s department. Particularly in dermatology

there is the possibility of reducing these waiting

periods to achieve a better time management. The

patient can photograph his skin lesions at any

location – either through a digital or mobile phone

camera, and then forward the provided image by

MMS or email to a dermatologist, who examines it.

Tele-dermatology has the potential to optimize

time management. Costs can be reduced in

comparison to the actual standard diagnostic way.

Efficiency of the outpatient clinics could be

increased with regard to economic considerations

and patient friendliness.

Recently published studies showed that cell

phones can be used in tele-dermatology. They are

very useful devices to produce photos easily. But

cell phones are not subject to medical device law. A

standard for produced images does not exist. So it

was necessary to test some cell phones and analyze

the produced images.

The impulse of this research was a tele-

dermatological pilot study at the Department of

Dermatology, Division of General Dermatology.

Photos were produced with cell phone cameras. The

major question was: Can camera phones produce

acceptable image quality for dermatology.

Before the study we analyzed three cell phones

with regard to quantitative and quality criteria.

1.1 Previous Work

Herrmann et al. got throughout positive results in a

study about tele-dermatological (TD) and face to

face (FTF) consultation. In a comparison of

diagnoses from 120 patients by tele-dermatologists,

agreement was from 46.4% respectively 70,2%

without additional information and 64,3%

respectively 76,6% with additional information. This

shows the absolute necessity of additional

information. Dermatologists felt certain with their

diagnoses in most cases. Sureness was evaluated

with a visual analogical scale from 0 to 10. Cases

difficult to assess with a tele-dermatologist were also

difficult for standard diagnoses and were therefore

identified with “doubtful diagnosis”. Furthermore,

image quality influenced the diagnosis. In 70% it

was possible to make a diagnosis. (Herrmann, 2005)

In 2004 Tai Khoa Lam tried to create photos

with two randomized mobile phones (Nokia 7650).

These photos were sent between a medical specialist

and an archivist. Photography was limited to the

hand trauma, radiographs or both. At the beginning

there was a discussion between the medical

specialist and the archivist who created a

management plan. After that, photos were sent via

mobile phone and multimedia. Now there was

another discussion and then the management plan

had to be modified. Within the following two

months 39 photos were sent. During the course of

the study there were four cases in which the

management plan had to be modified. (Lam TK,

2004)

Authors of „Telemedical Wound Care” devised a

study in which leg ulcers were photographed. 61 feet

were examined by three dermatologists. One of them

made FTF consultation and the other two were

responsible for mobile phones. They transmitted

images via e-mail. The result of the study was that

the image quality in 36 cases was “good” and in 12

cases “very good”. 50 of the involved parties felt

well and only one felt unwell. Three photos were of

poor-quality. (Braun RP, 2005)

In a cooperation of the Medical University of

Graz, Vienna and L’Aquila, diagnoses from camera

phone images and face-to-face dermatology were

compared. Two dermatologists examined these

images from 58 patients. In 48 patients a diagnosis

was provided. Six were immediately sent to a

dermatologist and four patients were advised to

come again a few days later.

During the following comparison of diagnoses in

41 cases, diagnosis was correct (full agreement). In

15 cases the diagnosis was wrong but still in the

same category of diagnoses (relative agreement). In

only three cases diagnosis was wrong

(disagreement). (Jauk B, 2006)

The Medical University of Graz (Austria) ran a

study researching the agreement between

teledermatology based on images from a cell phone

camera and face-to-face (FTF) dermatology. With a

quantity of 58 subjects two tele-dermatologists (TD)

analyzed the images produced. After checking, the

concordance between tele-diagnosis and FTF

diagnosis represented almost three-quarters (TD1:

71%, TD2: 76%). Nearly all diagnoses were in the

same diagnostic category (TD1: 97%; TD2: 90%).

(Ebner C, 2008).

APPLICABILITY OF MOBILE PHONES FOR TELE-DERMATOLOGY - A Pilot Study

475

2 MATERIAL AND METHODS

Image quality can be rated through different

methods. In this paper two kinds of analysis are

used. First an objective method with measurement

methods and second the subjective view of the

images.

2.1 Technical Methods

To acquire more information about colour fidelity,

illumination and MMS altering, some tests were

realized.

The tests involved three different cell phones.

Device 1 (D1) and Device 2 (D2) include a 3.2

megapixels camera. Device 3 (D3) includes a 5

megapixels camera and a “Tessar-lens” by order of

Carl Zeiss. These devices were selected randomly.

Tests were produced at the same settings and

conditions.

For photo analysis “Adobe Photoshop” was used

which produced histograms and relevant data.

Illumination was measured with “Minolta Auto

Meter IV F”. Scale unit is “exposure value” (ev)

which designates a number of combinations of glare

number (Relationship of the focal length f to the

diameter D of the entrance pupil) and exposure time,

which are equivalent to each other in photography

and photometry.

The colour fidelity test involves photographing a

single-coloured photo which was sent via mail.

Tolerance was 70. By the size of this parameter one

determines how large a colour area is in the selection

and the number of partly selected pixels increased or

reduced. (Selection range: 0-200)

Three series of tests were created at various

illuminations (5.3ev, 7.7ev and 9.1ev). Every

situation included 5 photos. The flash and the

camera light were always deactivated.

Every picture was analyzed with regard to

central pixel, mean pixel, standard deviation and

colour range. First a colour range was selected,

which includes the area with the most similar pixel.

In this area there is a central pixel and a mean pixel

with different standard deviations. All values are

derived from RGB spectrum.

2.2 Clinical Study

In a tele-dermatological pilot study at Department of

Dermatology, Division of General Dermatology,

Medical University of Vienna, 915 photos were

produced by 300 patients. Patients took photos of

their own dermatoses. If the dermatose was

unreachable, an assistant took the photo. Patients

only received a short instruction in the usage of the

mobile phone camera but they did not receive any

information how to obtain the best photo. Images

were created in a badly lit examination room without

any daylight.

3 RESULTS

3.1 Technical Results

The first test refers to colour fidelity. The following

three tables demonstrate the cumulative average

results of five photos per device. The “central pixel”

stands for the colour value which appears most in

contrast, “mean value” describes the average

brightness value. “Standard deviation” indicates how

strongly the brightness values vary. “Relative value”

is the amount of total pixel in percent. Central pixel

of template was 173 in RGB spectrum.

Table 1: Results at 5.3ev in a badly lit room.

central

pixel

mean

value

standard

deviation

relative

value

D1 150.80 151.35 23.74 94.44%

D2 83.20 90.80 22.17 99.03%

D3 140.80 139.52 13.97 73.69%

In Table 1 D1 and D3 nearly reached the original

pixel value. But D3 only includes 73.69% of the

mean value 139.52. D2 resulted in a mean pixel

value of 90.80 in an area of 99.03%.

Table 2: Results at 7.7ev with lateral illumination.

central

pixel

mean

value

standard

deviation

relative

value

D1 152.40 148.54 23.84 86.17%

D2 119.80 122.77 20.00 99.51%

D3 151.20 156.00 21.83 71.52%

With 7.7ev D1 and D3 reached a mean pixel value

of 148.54 (D1) or rather 156.00 (D3). D2 also

increased central pixel. It is about 122.77.

Table 3: Results at 9.1ev with vertical illumination.

central

pixel

mean

value

standard

deviation

relative

value

D1 150.60 148.95 24.14 88.91%

D2 77.80 101.97 44.44 90.57%

D3 160.00 163.37 16.46 56.48%

HEALTHINF 2009 - International Conference on Health Informatics

476

At maximum illumination (9.1ev) D1 reached a

mean value of about 148.95 (88.91% of image pixel)

and D3 reached 163.37 (only 56.48% of image

pixel). D2 reached a mean value of about 101.97.

Table 4: Average results at 5.3ev, 5.7ev and 9.1ev.

central

pixel

mean

value

standard

deviation

relative

value

D1 151,27 149,61 23,91 89,84%

D2 93,60 105,18 28,87 96,37%

D3 150,67 152,96 17,42 67,23%

In an area of 67.23% D3 took photos with a mean

standard deviation of 17.42. Mean pixel value was

nearly the original value. In contrast 96.37% at D2

had mean pixel of about 105.18. The average

standard was from 17.42 (D3) to 28.87 (D2).

Additional light sources can influence the

brightness distribution. In some areas in which light

reflect pixel, values are extremely falsified.

While sending MMS or email the original high

resolution photo stored on mobile phones will not

always remain unaffected. Usually, original images

were altered by lossy compression algorithms.

Images will be compressed, during sending via

MMS to a size of 640x480.

The following histograms present the

comparison between the original photo and

compressed photo after sending via MMS.

Figure 1: Histogram of original photo.

Figure 2: Histogram of MMS-photo.

Comparing the original photo with the MMS

photo, the histograms show that there is a minor

difference.

This reveals that colour distribution is nearly the

same. Mean pixel value changed from 185.63 to

183.89. After compression, central pixel value is 183

(before 185).

Image compression effects dimension and file

size. The following table shows the dimension and

file size of the images before and after sending via

MMS.

Table 5: Dimensions and file size.

dimensions file size

original MMS before after

D1 2048x1536 640x480 695 kb 13 kb

D2 2048x1536 640x480 748 kb 13 kb

D3 2592x1944 640x480 1852 kb 32 kb

3.2 Results of the Clinical Trial

Every phone produced about 22 blurred photos. In

total this results in only 66 photos.

15 photos were overexposed and therefore waste.

One of the three cell phones produced only one and

the worst only 8 overexposed photos.

4 DISCUSSION

Test results indicated that a homogeneously bright

illumination allows preventing errors and noise

artefacts. Diffuse illumination is the best scenario

for camera phone photos. Concentrated light creates

the most defects. It produces extremely bright pixels

in a wide area. Illumination had a considerably

unfavourable effect on colour fidelity.

While sending MMS, the original high resolution

photo stored on the mobile phone will not always

remain unaffected. Usually, original images are

altered by lossy compression algorithms.

The results of this pilot study show that the ability to

acquire close-up images and optimum illumination

are very important to obtain photos that are of

sufficient quality for tele-diagnosis.

REFERENCES

Herrmann, FE., Sonnichsen, K., Blum, A., 2005.

Teledermatology versus consultations--a comparative

study of 120 consultations. Hautarzt

Lam, TK., Preketes, A., Gates, R., 2004. Mobile phone

photo messaging assisted communication in the

assessment of handtrauma. ANZ J.Surg

Braun, RP., Vecchietti, JL., Thomas, L., 2005.

Telemedical Wound Care: Using a New Generation of

Mobile Telephones. Arch Dermatol

Ebner, C., Wurm, EM., Binder, B., Kittler, H., Lozzi, GP.,

Massone, C., Gabler, G., Hofmann-Wellenhof, R.,

Soyer, HP., 2008. Mobile teledermatology: a

feasibility study of 58 subjects using mobile phones. J

Telemed Telecare

APPLICABILITY OF MOBILE PHONES FOR TELE-DERMATOLOGY - A Pilot Study

477