Evaluation of Precision and Accuracy Assessment of Different 3-D Surface Imaging Systems for Biomedical Purposes

Abstract

Three-dimensional (3-D) surface imaging has gained clinical acceptance, especially in the field of cranio-maxillo-facial and plastic, reconstructive, and aesthetic surgery. Six scanners based on different scanning principles (Minolta Vivid 910®, Polhemus FastSCAN™, GFM PRIMOS®, GFM TopoCAM®, Steinbichler Comet® Vario Zoom 250, 3dMD DSP 400®) were used to measure five sheep skulls of different sizes. In three areas with varying anatomical complexity (areas, 1 = high; 2 = moderate; 3 = low), 56 distances between 20 landmarks are defined on each skull. Manual measurement (MM), coordinate machine measurements (CMM) and computer tomography (CT) measurements were used to define a reference method for further precision and accuracy evaluation of different 3-D scanning systems. MM showed high correlation to CMM and CT measurements (both r = 0.987; p < 0.001) and served as the reference method. TopoCAM®, Comet® and Vivid 910® showed highest measurement precision over all areas of complexity; Vivid 910®, the Comet® and the DSP 400® demonstrated highest accuracy over all areas with Vivid 910® being most accurate in areas 1 and 3, and the DSP 400® most accurate in area 2. In accordance to the measured distance length, most 3-D devices present higher measurement precision and accuracy for large distances and lower degrees of precision and accuracy for short distances. In general, higher degrees of complexity are associated with lower 3-D assessment accuracy, suggesting that for optimal results, different types of scanners should be applied to specific clinical applications and medical problems according to their special construction designs and characteristics.

Introduction

In the last few years, different three-dimensional (3-D) optical scanning systems have been applied in medicine to improve preoperative planning and surgical outcome research, most widely used in plastic and reconstructive surgery [1–4] and cranio-maxillo-facial surgery [5–7].

All 3-D optical scanning devices have in common that measurement is non-invasive, fast, non-deformable, with exact recording of targeted points in space of the captured body surface area, and the collected data can be processed employing ordinary personal computers using appropriate software solutions [8]. Most 3-D scanners rely on the technical principle of active or passive triangulation [9]. Active triangulation is based on a light source (laser light or fringe projection) and a detector positioned in a known distance. A beam projected onto the object is partly reflected and captured by the detector sensitive to the orientation of the incoming light. Since the angles of emission and incidence are known and the distance between emission and incidence is known, the location of all object points can be calculated in three dimensions. Photogrammetry is a passive triangulation system without the need of its own light source. In this approach, body surfaces are recorded by multiple cameras from different views. By identifying specific landmarks or markers on images taken from different angles, specific computer software can convert the data into a 3-D model using optical triangulation [9].

Before being applied to medical purposes, 3-D surface scanners were already used successfully for industrial purposes for quality control, construction and modification of industrial prototypes and engineered components [10, 11]. In the medical field, the established technology met new challenges: different anatomical regions of the human body feature structures of varying complexity. Therefore, standardization and validation studies of 3-D surface assessment were performed for different anatomical regions in previous studies [12–15], but comparison of different scanning systems for varying anatomical complexity are still lacking and limits characterization of medical application fields for specific 3-D scanning systems [16–18]. The fragmentary nature of the existing evidence and the fact that so far scanners have been tested mostly in small clinical series rather than in systematic and standardized studies, hinder assessment and further clinical dissemination [19]. The aim of the study is to comparatively evaluate the precision and accuracy of six 3-D imaging systems based on different scanning principles according to varying areas of anatomical complexity for biomedical purposes.

Materials and Methods

Research Objects

Five sheep skulls of different age (ranging from gestational age of 140 days to 2.5 months post-partum) were used as research objects after preparation and drying [20]. The sheep skulls present different areas of anatomical complexity defined as follows: Area 1 consists of the highly complex structures at the skull base; area 2 shows moderate structural complexity formed by the lateral skull sections, and area 3 presents low complexity with smooth surfaces at the top of the skull extending to the orbital region (Fig. 1). On every skull, 20 different anatomical landmarks were defined (Table 1), and 56 distance measurements (Fig. 1) between the landmarks were established as previously reported [20].

Fig. 1.

Sheep skulls from anterior, posterior, lateral and basal view (above left to right) with relating distance measurements according different areas of anatomical complexity: area 1 = high complexity (below right); area 2 = moderate complexity (below center); area 3 = low complexity (below left)

Table 1.

Definition of the 20 anatomical landmarks used for every sheep skull

Landmark	Definition
F	Spina nasalis caudalis
G	Processus pterygoideus, right side
H	Processus pterygoideus, left side
J	Processus muscularis of the bulla tympanica, right side
K	Processus muscularis of the bulla tympanica, left side
L	Incisura intercondylaris
M	Transition of the os occipitale to the condylus occipitalis, right side
N	Transition of the os occipitale to the condylus occipitalis, left side
O	Processus paracondylaris, right side
P	Processus paracondylaris, left side
Q	Arcus zygomaticus, right side
R	Processus lacrimalis caudalis, right side
S	Prominent bone in the upper orbital cavity, right side
T	Contact point of the processus frontalis of the os zygomaticus with the processus zygomaticus of the os frontale, right side
W	Arcus zygomaticus, left side
X	Processus lacrimalis caudalis, left side
Y	Prominent bone in the upper orbital cavity, right side
Z	Contact point of the processus frontalis of the os zygomaticus with the processus zygomaticus of the os frontale, left side
U	Contact point of the sutura cranialis of the os parietale with the os frontale
V	Contact point of the sutura cranialis of the os parietale

Correlation of Different Reference Measurement Methods

To analyze the measurement accuracy of the different optical 3-D surface imaging systems used in this study, the determination of a reliable reference measurement method is indispensible. Therefore, mean distance measurements between the above-described landmarks obtained by three reference methods including manual measurements, coordinate machine measurements and computer tomography measurements were compared with each other to investigate potential correlation.

Manual Measurements

Manual cephalometric measurements [millimeters] were taken five times on every skull by one examiner for 56 distances between the 20 defined landmarks within 24 h using a digital Digimatic 573 calliper (Mitutoyo GmbH, Neuss, Germany).

Coordinate Measurement Machine

The 3-D spatial coordinate values (x, y, z) for all 20 anatomical landmarks on each skull were determined five times by one examiner within 24 h using a computerized numerical control Zeiss UMC 850 coordinate measurement machine (Carl Zeiss IMT GmbH; Oberkochen, Germany). The provided spatial coordinates were converted into the 56 distance measurements by computing the distance between two landmarks A (x_a, y_a, z_a) and B (x_b, y_b, z_b) using the following equation in Excel 2007® (Microsoft Office 2007, Microsoft Deutschland GmbH, Germany)—distance A to B = √(x_a – x_b) ² + (y_a – y_b)² + (z_a – z_b)².

Computer Tomography

Computer tomography (CT) scanning of the five skulls was performed using a Siemens Somatom AR Star CT scanner (Siemens AG, Erlangen, Germany) at 110 kV and 120 mAs with a slice thickness of 1 mm and an isotropic resolution of 512 × 512 pixels at a color depth (greyscale) of 12 bits. Automatic segmentation, 3-D reconstruction and converting of the five skulls into virtual 3-D standard tessellation language (stl) models were performed with Mimics 13.0 (Materialise GmbH, Oberpfaffenhofen, Germany). Stl is a file format describing the surface geometry of a 3-D object without any representation of color, texture or other common computer-aided design (CAD) model attributes, and it is widely used for stereolithography-based rapid prototyping or other CAD applications. The 20 landmarks were placed five times on every virtual 3-D skull model using the RapidForm® 2002 PP1 SP1 software (INUS Technology, Inc., Seoul, South Korea) within 24 h by one examiner. The 56 defined distance measurements on each skull were computed according to the obtained spatial coordinate values on the 3-D skull model using the above-described equation in Excel 2007® (Microsoft Office 2007, Microsoft Deutschland GmbH, Germany).

Evaluation of 3-D Surface Imaging Systems

3-D Data Acquisition

Six different 3-D surface imaging systems were used in the study, all being non-invasive optical devices based on three different technologies (Table 2). The Vivid 910® (Minolta Co., Ltd., Osaka, Japan) and the FastSCAN™ (Polhemus Inc., Colchester, USA) devices are laser scanners; the PRIMOS® body and the TopoCAM® scanner (both GFMesstechnik GmbH, Teltow, Germany) as well as the Comet® Vario Zoom 250 (Steinbichler Optotechnik GmbH, Traunstein, Germany) are fringe light projectors. All these systems rely on the principle of active triangulation. The DSP 400® scanner (3dMD, Atlanta, USA) is optimized for medical applications and represents the passive triangulation concept of photogrammetry. The latter involves installation of four cameras facing the object in pairs of two from the left and the right side, which record an infrared spectral image of the surface surveyed. The image is an aid to recognize identical object points from different defined vantage points.

Table 2.

Manufacturers technical data of the scanners tested

	Method of measurement	Accuracy, mm	Scanning distance, mm	Scanning interval, s
Minolta Vivid 910®	Laser triangulation	0,068	Close-up, 600	2.5–0.3
Polhemus FastSCAN™	Laser triangulation	+/−1	150–200	30
GFM PRIMOS® body	Fringe light projection	≥0.03	Not specified	1.5
GFM TopoCAM®	Fringe light projection	0.025	1,600	≥20
Steinbichler Comet® Vario Zoom 250	Fringe light projection	+/−0.04	820	Not specified
3dMD DSP 400®	Photogrammetry	<0,5	1,000	0.0008

3-D Surface imaging of the sheep skulls was obtained five times with each of the six scanning systems in a standardized position using a mechanical rotary platform combined with a graduated disc to guarantee reproducible data acquisition. In order to capture a 360° view of the skulls, 18 single shots were needed for each skull. Exceptions were the Comet® Vario Zoom 250 innately equipped with the Comet® Rotary computer-guided rotary platform and the FastSCAN™. The FastSCAN™ is a handheld laser scanner with a magnetic tracking system to automatically register and convert the single scans into one virtual 3-D model in real time, which saves time-consuming post-processing of the data.

3-D Data Processing and Analysis

The 18 single shots of each acquisition were converted into one single 3-D skull model by identifying corresponding points on each image using the RapidForm® 2002 PP1 SP1 software (INUS Technology, Inc., Seoul, South Korea). During the 3-D processing of the single shots, we deliberately avoid any data optimization to ensure faithful 3-D data assessment. On every virtual 3-D skull model, one observer selected five times within 24 h the above-defined 20 landmarks using the RapidForm® software. The spatial coordinate values for each landmark on each 3-D skull model were collected, and the 56 distance measurements were computed according to the above-described equation in Excel 2007® (Microsoft Office 2007, Microsoft Deutschland GmbH, Germany). Furthermore, precision and accuracy of the six 3-D scanning assessments were comparatively evaluated according to the different areas of complexity and the measured distance length.

Statistical Analysis

Replicate distance measurements in millimeters for each reference method and scanner used were aggregated as the mean and standard deviation (SD). Pearson correlation coefficients (r) were used to investigate the linear relationship between distance measurements of the three reference methods. Statistical analysis including descriptive statistics, correlation and regression analysis, paired-sample t test for control sample comparison, error bars and scatter-plots was carried out using SPSS® version 11.5.1 for Windows (SPSS Inc., Chicago, IL, USA). All tests were two-tailed using a global significance level of p < 0.05.

Results

Correlation of Different Reference Measurement Methods

High correlations were observed between all the three reference methods (Table 3 and Fig. 2), showing highest agreement between CMM and CT measurements (r = 0.998; p < 0.001). Manual measurements correlated equally with the other methods (for both r = 0.987; p < 0.001). According to the different areas of anatomical complexity, manual measurements in area 3 with low complexity revealed lower correlations with CMM measurements (r = 0.903, p < 0.001) as well as with CT measurements (r = 0.894, p < 0.001). Regression analysis showed no statistical differences (p = 0.161–0.955) between the standard deviations of manual measurements for the areas of varying complexity in dependence on the distance length (Fig. 3). Because no obvious statistical difference occurred to the other methods and manual measurements are applied in all day clinical routine, it was therefore appropriate to use the manual measurements as the reference method for the following precision and accuracy assessment of the 3-D measurements.

Table 3.

Pearson correlation coefficients (r) analysis to investigate the linear relationship between distance measurements of the three reference methods using a global significance level of p < 0.05: manual measurement (MM), coordinate machine measurements (CMM) and computer tomography (CT) measurements

	MM				CT
	Area 1	Area 2	Area 3	Mean	Area 1	Area 2	Area 3	Mean
CT	r = 0.999	r = 0.999	r = 0.894	r = 0.987
	For all p < 0.001
CMM	r = 0.999	r = 0.999	r = 0.903	r = 0.987	r = 0.998	r = 0.998	r = 0.994	r = 0.998
	For all p < 0.001				For all p < 0.001

Fig. 2.

Correlation analysis between different reference methods. Manual measurement (MM), coordinate machine measurements (CMM) and computer tomography (CT) measurements. The dots represent the different distance measurements of the three reference methods used for the correlation analysis

Fig. 3.

Regression analysis between the standard deviations of manual measurements for the areas of varying complexity in dependence on the distance length in millimetres showing no statistical differences (p = 0.161–0.955)

Precision Evaluation of 3-D Scanning Assessment

The precision of the 3-D scanning assessment is expressed as the mean standard deviation of multiple measurements carried out between the 20 landmarks. The mean measurement precision according to the areas of complexity obtained by the six different scanners, and manual measurements are summarized in Table 4.

Table 4.

Precision assessments for all scanners and manual measurements according to the areas of complexity expressed as the mean standard deviation in millimetres of multiple measurements carried out between the 20 landmarks

Evaluations	Area 1 SD	Area 2 SD	Area 3 SD	All areas SD
Precision evaluation, SD (mm)
Minolta Vivid 910®	0.45	0.29	0.25	0.33
GFM TopoCAM®	0.38	0.26	0.22	0.29
GFM PRIMOS® body	0.53	0.4	0.37	0.43
Steinbichler Comet® Vario Zoom 250	0.41	0.26	0.22	0.30
Polhemus FastSCAN™	0.92	1.135	0.86	0.97
3dMD DSP 400®	0.71	0.565	0.43	0.57
Manual	0.21	0.21	0.20	0.21
Accuracy evaluation, mean manual − mean scanner (mm)
Minolta Vivid 910®	−0.34	−0.26	0.05	−0.18
GFM TopoCAM®	−0.98	−0.49	−0.30	−0.59
GFM PRIMOS® body	−0.71	−0.26	0.05	−0.31
Steinbichler Comet® Vario Zoom 250	−0.37	−0.20	−0.08	−0.22
Polhemus FastSCAN™	−1.10	−1.25	−2.13	−1.49
3dMD DSP 400®	0.36	−0.19	−0.68	−0.17

Accuracy assessments for all scanners expressed as the mean calculated difference between manual measurements (reference method) and the corresponding mean scanner measurements in millimetres for the different areas of complexity

Irrespective of the structural anatomical complexity of the surfaces, the TopoCAM® and the Comet® Vario Zoom 250 showed the highest precision. Precision with the other scanners was acceptable in all areas and all skulls, ranging from 0.33 to 0.57 mm, except for the FastSCAN™ which produced a mean standard deviation over all areas of 0.97 mm (p > 0.05). Furthermore, area complexity and scanner precision correlated: The more complex the analyzed area, the lower the degree of scanner precision. The FastSCAN™ represented an exception with the lowest precision in the area of moderate complexity (area 2). In the area of highest complexity (area 1), the TopoCAM® (SD = 0.38 mm) was slightly more precise than the Comet® Vario Zoom 250 (SD = 0.41 mm).

Scanning precision assessment in accordance to the measured distance length revealed that most 3-D devices present higher measurement precision for large distances and lower degrees of precision for short distances. One exception to this finding is again the FastSCAN™ showing highest measurement precision for short distances (Table 5).

Table 5.

Precision assessment expressed as the mean standard deviation in millimetres for all scanners and manual measurements in accordance to the measured distance length in centimetres

System	SD (mm) distance length, 0–4 cm	SD (mm) distance length, 4–7 cm	SD (mm) distance length, 7–10 cm
Minolta Vivid 910®	0.42	0.36	0.24
GFM TopoCAM®	0.32	0.27	0.24
GFM PRIMOS® body	0.5	0.35	0.29
Steinbichler Comet® Vario Zoom 250	0.3	0.25	0.25
Polhemus FastSCAN™	0.91	0.94	1.11
3dMD DSP 400®	0.53	0.52	0.47
Manual	0.21	0.19	0.19

Accuracy Evaluation of 3-D Scanning Assessment

The accuracy of the 3-D scanning assessment is expressed as the mean calculated difference between manual measurements (reference method) and the corresponding mean scanner measurements (manual mean MM − scanner mean SM). The mean measurement accuracy for the different areas of complexity obtained by the six different scanners is summarized in Table 4.

Regardless of the complexity of the anatomical areas, the Vivid 910®, the Comet® Vario Zoom and the DSP 400® produced the most accurate assessments (Table 4 and Fig. 4). The Vivid 910® showed the highest mean accuracy in the most complex area (area 1), and in the area with low complexity (area 3), the DSP 400® showing the highest mean accuracy of all scanners in area 2 slightly followed by the Comet Vario® Zoom (Table 4). Significant differences were assumed for non-overlapping confidence intervals (Fig. 5). Accuracy was significantly different with the FastSCAN™ showing less accurate results compared with all the other scanners tested in all areas of complexity (p > 0.05). In general, higher degrees of complexity are associated with lower 3-D assessment accuracy (Fig. 5).

Fig. 4.

Accuracy assessment expressed as the difference between manual measurements and the computed scanner measurements for all scanners regardless of the complexity of the anatomical areas

Fig. 5.

Accuracy assessment expressed as the difference between manual measurements and the computed scanner measurements for all scanners according to the areas of varying structural complexity

Furthermore, the accuracy of the scanner measurements with respect to the measured length of all distances obtained by manual measurements was evaluated by regression analysis (Fig. 6). Best agreement for the overall length was noticed for the DSP 400®, Vivid 910® and the Comet Vario®. Again, it becomes evident that the FastSCAN™ delivered higher mean distances than manual measurements. Except the DSP 400® and the FastSCAN™, all scanners showed larger difference measurements for short distances, with decreased differences for long-distance lengths.

Fig. 6.

Accuracy assessment for all scanners expressed as the difference between manual measurements and the computed scanner measurements in dependence on the distance length in millimetres

Discussion

This study aimed to evaluate the accuracy and precision assessment of different optical 3-D scanning systems based on diverse physical principles for biomedical applications on a static biological model consisting of surface structures of varying complexity. Evidently, not all scanning systems available on the market could be included in our series. The sheep skulls have been chosen as research object because manual distance measurements between well-established anatomical landmarks has been approved and validated in our previous study [20]. With human subjects, reliable measurements exploring limitations in accuracy and precision of the systems would have hardly been possible, given a substantial amount of variables that cannot be standardized, such as skin translucency, reflectivity, colour or artifacts caused by movement [13–18]. So far, a number of publications used single scanners for anthropometric measurements or clinical studies of individual body regions [21–26].

When comparing scanner results with an established reference method (e.g. manual measurements), distinction has to be made between accuracy and precision and should be considered independently [13–15]. The term accuracy describes how close results come in numerical terms to the reference method; precision gives the amount of variation or “scattering” of the results around a given reference. The ideal scanner should feature both: high accuracy and precision. Studies evaluating scanners in medicine have mostly investigated one special anatomical area with a particular scanner and found accuracy and precision of the 3-D devices reasonable enough to recommend them for clinical application without investigating if the specific scanning systems or maybe a different device is best suited for the anatomical region of interest [12–26]. Our study, testing various scanners based on different technologies on three different kinds of complex surfaces for the first time, is able to corroborate these findings. Therefore, the sheep skull values should be taken as examples of a rough sensor performance overview achieved under normal clinical conditions by an experienced medical staff and not as the real precision or accuracy values documented by the manufacturer. According to our previous studies which showed that landmark positioning on 3-D models are sufficiently precise and accurate, resulting in observer independent, reproducible measurements [3, 13–15, 20, 25] by also analysing the influence of various experimental conditions (as lighting and number of scanners, alignment of laser stripe light and number of scanner shots, angle of recording, position of the region of interest during acquisition, premarking of anatomical landmarks) on the precision and accuracy of 3-D measurements [13–15], the presented study abstains from further intra- and inter-observer variability analysis regarding measurement precision and accuracy. Furthermore, as manual measurements are currently still applied in all day clinical routine as the standard reference method, the presented study used these manual measurements as the reference method for precision and accuracy assessments. However, this appropriate approach could have been additionally refined by superimposing the different 3-D surfaces of the sheep skull models on each other and by measuring the 3-D surface contour difference between the models in order to get an overall idea concerning potential 3-D surface acquisition divergences. 3-D Surface contour analysis should be favoured, and future studies are under way, but have not been considered as an additional analysis in this study because of manuscript clarity, readability and comprehensibility.

Overall, the Comet® Vario Zoom showed the best results to accuracy and third best to precision. These test results qualify the Comet® Vario Zoom to be used in various biomedical applications involving complex biomedical surfaces. Due to the size of the sheep skulls in our study, we used the computer-guided rotary table Comet® Rotary as accessory equipment to the Comet® Vario Zoom. Other manufacturers (e.g. Konica, Minolta) offer similar accessories to their scanners; these, however, were not tested in our study. Combinations of high-accuracy scanning systems with such computer-guided rotary tables might be well suited for dental applications or forensic applications.

Interestingly, the Comet® Vario Zoom and the Minolta Vivid 910® scanners produced overall very similar results in our study, although relying on different technologies and despite the fact that one device was combined with a rotary table while the other was not. This indicates that the computer-guided rotary table enables quick automatic data recording but has no influence on accuracy or precision. When 18 single-scanner camera shots are needed as in our series, the manual set-up requires lots of time.

Advantages of the Vivid 910® scanner are good quality of recording, high accuracy and good precision, compact housing and easy operation. With different optical lenses available, the scanner is capable to adapt easily to surface areas of different size or volume that are to be surveyed. The procedure of calibrating the instrument before the start of the recording session, common in the operation of many scanners, is not necessary with the Vivid 910®. The presented study indicates a high level of agreement to our previous clinical study findings with the Vivid 910® [13–15]: In anatomical areas with higher complexity (mouth, nose, submammary region), the scanner precision is lower compared with areas with moderate or low anatomical complexity (eye, cheek area, thoracic region). With the exception of the FastSCAN™, precision of the scanners tested deteriorated when surface structures became more complex, a potential biomedical application field for this scanner. Good results in the survey of complex surfaces of this handheld linear laser scanner obtained on difficult surface structures benefit from the fact that the targets can be monitored very flexibly from different directions and angles.

The PRIMOS® body ranked in the middle of our study. It is worth highlighting the very good texture information of the captured surfaces of the PRIMOS® body and its automatic registration of the single surface shells.

The accuracy of the DSP 400® scanner and its midfield precision ranking (wide variation of recorded data about the mean) also necessitate a comment focusing on fundamental methodical characteristics of photogrammetry as described above. The results are best, when the captured surfaces feature many individual structures [8]. This finding explains why the DSP 400® scanner performed best in the complex anatomical area with many distinctive landmarks and features. Also, with this scanner, the phenomenon of “shadow casting” is of minor importance than with laser or fringe light projectors because it does not use an own light source. Maybe this technical aspect explains why photogrammetry is primarily used for facial acquisition and only found potential application in breast region acquisition after technical improvements.

A 360° assessment of larger parts or the whole surface of the human body and larger biological bodies are very important in a broad number of biomedical applications. Unfortunately, it is hardly possible to immobilize the patient completely for such an examination. For the recording of large surfaces, such as the whole human body, custom-made scanning systems are required. Companies like Cyberware® Incorporated (Monterey, CA, USA), InSpeck® Inc (Montreal, Quebec, Canada) or Human Solutions® GmbH (Kaiserslautern, Germany) develop solutions for such purposes and have to be tested in the future for biomedical purposes.

No scanner in our series and, correspondingly, no underlying technology or scanning systems proved to be superior in monitoring the static biological surfaces in our study. Different performances were mainly related to constructional aspects of the devices tested. Evidently, the wide variety of possible biomedical applications challenges scanner technology in manifold ways and with special requirements. Scanners for biomedical purposes must be designed to suit specialties from dentistry, which necessitates accuracy in the range of 10⁻¹–10⁻³ cm to plastic and reconstructive surgery, where whole-body surveys can easily do with accuracy in the range of 1 or 2 mm. Potential human artefacts such as movements, change of position, breathing or unusual anatomical characteristics places enormous demands on the 3-D scanning devices in terms of reducing the acquisition time. The clinical necessity accelerated the development of faster devices (acquisition time under 0.5 s), but still acquisition time significantly diverges between the different scanning systems from minutes (FastSCAN™), over a view seconds (Minolta Vivid 910®) to milliseconds (3dMD DSP 400®). By applying a standardized 3-D scanning protocol, the influence of human artefacts can be compensated and reduced [13–15, 23–26]. 3-D imaging of the breast region for example can be obtained in a standing position on predefined markers on the ground, patients back-supported by a wall, patients holding their breath during acquisition and arms down the side crossed behind at the height of the pelvis [3, 15, 23, 25]. In addition, Patete et al. presented a methodology for active compensation of breathing motion and involuntary movements to obtain a reliable and artefact-free patient’s body surface during the 3-D acquisition [27]. Further developments in the field of holography may play an important role in recording the human body very precisely and in true colour images in very short time [28]. Therefore, a global recommendation for a specific 3-D scanning system is nearly impossible and has to be adapted to the field of biomedical application. We therefore believe that in the near future 3-D scanning systems will be produced according to the specific anatomical requirements and will be no more provided as a general solution for all biomedical purposes and anatomical varieties.