**2.2. Model configurations and associated protocols**

As mentioned above, in order to analyse the full human skeleton posture and movement, our group started a project in the mid‐1990s to create a complete and accurate 3D biomechanical model of the human skeleton, with particular attention to the accurate reproduction of spinal detail. This was achieved by merging different segmental biomechanical models presented in the literature [7].

The accuracy and precision of the model rely on anatomical findings (cadaver dissections, in vivo X‐ray and gamma ray measurements parametric regression equations) [32–35] as well as a variety of signal‐processing procedures and optimization methods described in the lit‐ erature [6–8, 27–31]. The model was conceived in an adjustable form in order to enable the scaling of the individual characteristics of any subject. This was achieved by fitting the 3D anthropometric size of a given skeletal segment, to 3D opto‐electronic measurements of the corresponding body landmarks labelled by passive retro‐reflective markers [7].

The model can work at different levels of complexity depending on the purposes and require‐ ments of the analysis. In fact, when necessary, it is possible to increase the level of detail for parametric scaling by obtaining additional anthropometric measurements relative to specific anatomical segments. To this end, various protocols involving different body labels have been established for different analytical purposes [6–8, 27–31].

The protocol, which has been developed for the analysis of human posture and spine‐related pathologies (scoliosis, back pain, etc.), is based on 27 selected human body bony prominences accurately identified by palpation. The comprehensive list of these anatomical landmarks is given in **Figure 2**. An example of the related full 3D skeleton reconstruction, obtained via opto‐electronic stereo‐photogrammetric measurement, can be found in **Figure 3**.

For gait analysis, a 49‐marker protocol has been set‐up to appropriately define the pelvis‐ lower limb kinematic chain [8, 27, 28] (**Figure 4**).

The posture and gait analysis protocols share the same landmarks that are set for the head, trunk and the pelvis. Within the 49‐marker protocol, the pelvis‐lower limb apparatus is mod‐ elled by a 9‐link chain model, with the pelvis, thigh, shank and foot links joined by ball and socket joints representing the hips, the knees and the ankles, respectively. At least three mark‐ ers per each segment have been used, in particular: for the pelvis: (as above) bilateral anterior and posterior superior iliac spine (ASIS and PSIS) bony landmarks; femur: great trochanter A 3D Spine and Full Skeleton Model for Opto-Electronic Stereo-Photogrammetric Multi-Sensor... http://dx.doi.org/10.5772/intechopen.68633 25

For posture evaluation (six IR TVCs at 0.3 Mpix resolution), the usual acquisition volume (i.e. the physical calibrated volume of the room, inside which the subject can be measured with known accuracy and precision) is in general somewhat like 3‐m wide by 3‐m deep by 2‐m high. With such a configuration, the usual final mean 3D stereo‐photogrammetric error is limited to a range of 0.3–0.4 mm throughout the entire working volume. If higher resolution IR TVCs are used, the 3D stereo‐photogrammetric error is even lower. The whole calibration phase takes less than 5 min. This calibration step is needed only when the cameras are initially installed or when they are moved to a new position. When the installation is fixed in a research laboratory,

the calibration is performed only occasionally to maintain an accurate calibration level.

As mentioned above, in order to analyse the full human skeleton posture and movement, our group started a project in the mid‐1990s to create a complete and accurate 3D biomechanical model of the human skeleton, with particular attention to the accurate reproduction of spinal detail. This was achieved by merging different segmental biomechanical models presented in

The accuracy and precision of the model rely on anatomical findings (cadaver dissections, in vivo X‐ray and gamma ray measurements parametric regression equations) [32–35] as well as a variety of signal‐processing procedures and optimization methods described in the lit‐ erature [6–8, 27–31]. The model was conceived in an adjustable form in order to enable the scaling of the individual characteristics of any subject. This was achieved by fitting the 3D anthropometric size of a given skeletal segment, to 3D opto‐electronic measurements of the

The model can work at different levels of complexity depending on the purposes and require‐ ments of the analysis. In fact, when necessary, it is possible to increase the level of detail for parametric scaling by obtaining additional anthropometric measurements relative to specific anatomical segments. To this end, various protocols involving different body labels have been

The protocol, which has been developed for the analysis of human posture and spine‐related pathologies (scoliosis, back pain, etc.), is based on 27 selected human body bony prominences accurately identified by palpation. The comprehensive list of these anatomical landmarks is given in **Figure 2**. An example of the related full 3D skeleton reconstruction, obtained via

For gait analysis, a 49‐marker protocol has been set‐up to appropriately define the pelvis‐

The posture and gait analysis protocols share the same landmarks that are set for the head, trunk and the pelvis. Within the 49‐marker protocol, the pelvis‐lower limb apparatus is mod‐ elled by a 9‐link chain model, with the pelvis, thigh, shank and foot links joined by ball and socket joints representing the hips, the knees and the ankles, respectively. At least three mark‐ ers per each segment have been used, in particular: for the pelvis: (as above) bilateral anterior and posterior superior iliac spine (ASIS and PSIS) bony landmarks; femur: great trochanter

corresponding body landmarks labelled by passive retro‐reflective markers [7].

opto‐electronic stereo‐photogrammetric measurement, can be found in **Figure 3**.

**2.2. Model configurations and associated protocols**

24 Innovations in Spinal Deformities and Postural Disorders

established for different analytical purposes [6–8, 27–31].

lower limb kinematic chain [8, 27, 28] (**Figure 4**).

the literature [7].

**Figure 2.** Protocol for 3D posture analysis: list of 27 anatomical landmarks identified by palpation.

**Figure 3.** Subject's pictures and assessed full 3D skeleton reconstruction.

middle thigh and lateral and medial femoral epicondyles; shank: head of fibula, tibial tuber‐ osity, middle shank, lateral and medial malleoli; foot: heel (calcaneus process), the first and fifth metatarsal heads, distal big toe end point (**Figure 4**).

**Figure 4.** Protocol for 3D gait analysis includes a set of 49 identified anatomical landmarks.

The foot marker set in particular allows for the modelling of the foot and is subdivided into two different rigid bony segments: fingers‐forefoot segment identified by big toe first and fifth metatarsal heads and rear‐mid‐foot identified by the first and fifth metatarsal heads and heel (see the small panel representing the flexed forefoot in **Figure 4**). This latter choice allows for a more accurate description of the ankle‐foot complex biomechanics. By using regression equa‐ tions [32], the ASIS and PSIS positions provide the basis for the assessment of hip joint centre positions and of pelvis width. The knee joint centre is taken as the mid‐point of the segment linking the femur medial and lateral epicondyles, and the ankle joint centre as the mid‐point of the segment joining the two malleoli. A modified version of this latter protocol (adding the humerus lateral epicondyle and ulnar styloid landmarks) allows the determination of the upper limb positions when they are of interest.

When the focus of the analysis is on neck and back pain, specific test batteries have been estab‐ lished to evaluate the related postural and spinal dysfunction. In this case, a further three‐ marker set is placed on a head band and are added to either the 27‐ or 49‐marker protocols in order to be able to reconstruct the head and neck even during a forward‐bending test. In fact, during the patient's forward‐bending execution the three markers placed on the face (zygo‐ matic bones and chin needed to measure the skull position and orientation) usually disappear from the TV‐cameras' field of view. In this case, the skull landmarks cannot be identified anymore. To overcome this problem, three added markers are placed on a head band: during orthostatic posture acquisition, the rigid geometrical relationship occurring between the head band markers and the anatomical skull markers ones is established.

Afterwards, during forward bending, the markers on the head band always remain visible to the cameras, thus allowing the skull and neck position and orientation reconstruction alongside all the movement measurements. In the same way, head axial rotations can also be evaluated for neck pain patients. In addition to the 3D kinematic measurements, the GOALS system capability can be expanded, when necessary, by gathering additional biomechanical data (forces, pressures, electromyographic signals, etc.) measured simultaneously by different sensors, allowing to per‐ form what is defined in the literature as a 'multi‐factorial approach' [36] (**Figure 5**).

The 'multi‐factorial approach' means the capability to fully integrate all these measurement data into a unified approach to correlate and combine all the morphological‐kinematic char‐ acteristics measured in order to achieve a full‐functional evaluation with the aim of using the results of such measurement for clinical purposes and objectives.

By using the multi‐factorial approach, additional useful functional information is available. In particular, by using ground reaction forces and the segmental inertial and gravitational contributions derived from the skeleton/body model, it is possible to assess the joint net forces and torques at each lower limb joint and even at each spine inter‐vertebral level by using a model derived from Liu and Wickstrom in 1973 [35].

Baropodometric platforms and/or baropodometric in‐shoe insole system allow to measure the underfoot pressure distribution maps. These latter are very useful when a better descrip‐ tion of the foot‐floor interaction is needed to proceed to plantar foot orthoses custom design.

When of interest, SEMG is recorded by a telemetric system following the SENIAM European project recommendations [37]. In gait analysis, the activity of lower limb muscles is recorded to investigate motor co‐ordination/dysfunction. In low back pain (LBP), the multifidus (MF) and erector spinae‐longissimus dorsi (ESLD) activities are bilaterally collected to study the flexion‐relaxation phenomenon (FRP) (see case n. 4 in Section 3 for a detailed explanation) [38]. In neck pain patients, SEMG is generally recorded bilaterally on the upper trapezius and sternocleidomastoid muscles.

The foot marker set in particular allows for the modelling of the foot and is subdivided into two different rigid bony segments: fingers‐forefoot segment identified by big toe first and fifth metatarsal heads and rear‐mid‐foot identified by the first and fifth metatarsal heads and heel (see the small panel representing the flexed forefoot in **Figure 4**). This latter choice allows for a more accurate description of the ankle‐foot complex biomechanics. By using regression equa‐ tions [32], the ASIS and PSIS positions provide the basis for the assessment of hip joint centre positions and of pelvis width. The knee joint centre is taken as the mid‐point of the segment linking the femur medial and lateral epicondyles, and the ankle joint centre as the mid‐point of the segment joining the two malleoli. A modified version of this latter protocol (adding the humerus lateral epicondyle and ulnar styloid landmarks) allows the determination of the

**Figure 4.** Protocol for 3D gait analysis includes a set of 49 identified anatomical landmarks.

When the focus of the analysis is on neck and back pain, specific test batteries have been estab‐ lished to evaluate the related postural and spinal dysfunction. In this case, a further three‐ marker set is placed on a head band and are added to either the 27‐ or 49‐marker protocols in order to be able to reconstruct the head and neck even during a forward‐bending test. In fact, during the patient's forward‐bending execution the three markers placed on the face (zygo‐ matic bones and chin needed to measure the skull position and orientation) usually disappear from the TV‐cameras' field of view. In this case, the skull landmarks cannot be identified anymore. To overcome this problem, three added markers are placed on a head band: during orthostatic posture acquisition, the rigid geometrical relationship occurring between the head

Afterwards, during forward bending, the markers on the head band always remain visible to the cameras, thus allowing the skull and neck position and orientation reconstruction alongside all the movement measurements. In the same way, head axial rotations can also be evaluated for neck pain patients. In addition to the 3D kinematic measurements, the GOALS system capability can be expanded, when necessary, by gathering additional biomechanical data (forces, pressures,

upper limb positions when they are of interest.

26 Innovations in Spinal Deformities and Postural Disorders

band markers and the anatomical skull markers ones is established.

**Figure 5.** General 'multi‐factorial‐multi‐sensors' experimental set‐up for GOALS system for 3D posture and movement analysis.
