A practical checklist for front and right-side photos
To make posture photos easier to compare, keep the conditions consistent: camera position, distance from the person, viewing angle, lighting, and the instructions given before each shot. Include the whole body in the frame and make sure the anatomical landmarks remain clearly visible.
Two photos of the same person can look different because the phone is tilted or the torso is slightly rotated. These differences also affect the measurements. Check the photography setup before uploading an image to the app. [1]
This protocol is intended for professionals documenting standing posture. It will help you organize photography in your practice or at home and prepare photos for PAX. Other positions, such as forward bending, require separate instructions.
Quick checklist before taking photos
- One person is standing on a level surface, with a background that contrasts with their clothing and body.
- The phone is secured; the lens, zoom setting, height, and distance have been selected and recorded.
- The camera is aimed horizontally, without tilting to either side.
- The head, hands, and feet are fully within the frame, with space around the body.
- Clothing and hair do not hide the landmarks needed for the selected analysis.
- The person has adopted the agreed stance and is not deliberately trying to “correct their posture.”
- They are looking straight ahead at eye level, rather than down at the phone.
- The image is sharp, is not mirrored, and has not been edited to change the shape of the body.
- For a follow-up assessment, the previous conditions have been reproduced and the same analysis mode selected.
The sections below explain each step. The suggested camera heights and distances are practical starting points. They are not a universal medical standard or a validated protocol establishing the accuracy of PAX.
1. Set up a consistent photography space
Choose an area with a level floor and a plain, uncluttered background. The person should fit in the frame from head to toe, with the camera far enough away that you do not need to tilt the phone.
Lighting should make the body contours and relevant landmarks visible. A bright window behind the person can turn them into a dark silhouette. Strong lighting from one side can leave some landmarks in shadow. Take a test photo and zoom in to check that the body contours and relevant landmarks are clearly visible.
Mark the tripod position and the line where the person will stand. Separate foot-position marks for the front and right-side views can be useful. At the next visit, these marks will help you recreate the setup without estimating it by eye.
The marks help reproduce the chosen foot position in later photos.
Do not ask the person to press the back of their head, shoulder blades, or pelvis against a wall when the aim is to record their habitual stance. That instruction creates a different body position. Leave enough room for them to stand comfortably and turn around.
2. Secure the camera and choose the distance
Use camera settings you can reproduce
Secure the phone on a tripod or stable mount. For a straightforward setup, use the main rear camera at its standard zoom setting, usually 1×. Use the same lens and zoom setting at subsequent assessments.
If the person does not fit in the frame, first move the camera farther away. Turn off filters, body reshaping, and other features that could alter the image's geometry.
How high should the camera be?
For a full-body photo, a practical starting point is to place the lens at about half the person's height. For someone who is 180 cm tall, that is approximately 90 cm above the floor. Measure to the lens itself, not to the top edge of the phone.
Aim the camera horizontally and position the person in the center of the frame. Do not tilt the phone upward to fit in the head or downward to include the feet. Adjust the distance and check the framing again instead.
How far away should the camera be?
For the initial setup, try a distance of about 2.5 m from the lens to the line where the person stands. Then adjust it to suit the person's height, the camera, and the available space. The whole body should fit comfortably within the frame, with the relevant details still clearly visible. Record the actual distance you choose.
Mota and colleagues compared photographs of a mannequin taken from 3 and 5 m, with the camera positioned at half the mannequin's height. Other research protocols use different settings. These findings help identify the conditions that need to be controlled, but they do not establish one mandatory distance for every smartphone and person. [2]
If you need to use a different phone or move the camera for follow-up photos, record the change.
An example of a smartphone mounted on a tripod. Record the chosen lens height and distance so you can reproduce them at the next session.
3. Choose clothing that leaves landmarks visible
Choose clothing that makes the relevant body areas visible without creating a misleading outline. Shorts and a close-fitting top are often practical. The exact choice depends on the landmarks being used: a waistband may obscure the pelvic area, while a loose T-shirt can hide the position of the shoulders and trunk.
An example of a full-body front photo with space around the person.
Move hair away from the ears and neck, and remove bulky items from pockets. Keep footwear conditions consistent throughout a series of photos. If the initial photos were taken barefoot, take the follow-up photos barefoot too. Photographing someone in shoes may be a separate task, but do not mix those images with barefoot photos without recording the difference.
Do not use a waistband to identify pelvic position or a T-shirt seam to locate the shoulder. If a required anatomical landmark is not visible, adjust the preparation for the photo or acknowledge the limitation when placing the point.
In the SAPO study, agreement between measurements varied across parameters even when the same photographs were marked again. Check the points systematically and use consistent rules for identifying them. [4]
4. Give the same instructions before each photo
Decide in advance which stance you want to record. For habitual standing posture, a short instruction can be enough:
“Stand comfortably on the marked spot. Let your arms hang loosely. Look straight ahead at eye level. Breathe normally.”
Then allow a few seconds for the person to stand quietly. Do not add “pull your shoulders back,” “pull your stomach in,” “lift your chest,” or “stand up straighter”: these instructions change the task. If you also want a photo after active postural correction, label it as a separate condition.
Choose a comfortable foot position and record it for follow-up photos. There is no need to force the heels together, turn the toes to a prescribed angle, or impose symmetry. If you are using a specific protocol with a defined stance width, apply it consistently and record that choice.
Keep the arms in the agreed position. If a particular measurement requires moving them slightly away from the body, follow the same method throughout the series. Changing arm position arbitrarily between visits makes comparison more difficult.
Pay particular attention to gaze direction. A camera positioned at half the person's height is below their face. Asking them to “look into the lens” may make them lower their head. Instead, ask them to look straight ahead at eye level. If necessary, use a visual target at that person's eye height.
Stolinski and colleagues standardized the positions of the feet and arms and the direction of gaze, distinguishing habitual stance from actively corrected posture. The study involved children aged 7–10 years, so its specific settings and reference values cannot automatically be applied to adults. The useful principle here is to give consistent instructions. [3]
5. Take front and right-side photos
PAX's Quick Analysis mode uses front and right-side photos. “Right side” means the right side of the person being photographed; mirroring a left-side photo does not make it a right-side view. [7]
An example of a right-side photo: the person's own right side faces the camera.
| What to check | Front view | Right-side view |
|---|---|---|
| Position relative to the camera | The person faces the camera without turning partly to one side. | The person's right side faces the camera; they turn their whole body. |
| Head and gaze | Looking straight ahead at eye level. | Looking straight ahead relative to the body, without turning the head toward the photographer. |
| Arms | Hanging loosely unless the selected method requires a different position. | The position is reproduced consistently, with the required landmarks visible. |
| Framing | The whole head, hands, and both feet are included, with space around the body. | The whole head, hands, heels, and toes are included, with space around the body. |
When changing views, ask the person to turn their whole body and stand on the corresponding marks. Keep the camera in position and check that the person remains on the chosen photography line.
For the front photo, do not deliberately level the shoulders or pelvis. Check how the person is oriented toward the camera while leaving their individual body alignment as it is.
Diagram of the views and camera setup. D is the chosen distance to the photography position; h is the lens height. The figure is schematic and does not represent an ideal posture.
If Full Analysis mode is available in your version of PAX, follow its prompts for additional views.
6. Check image quality before finishing
Open the photo and zoom in. Check that the important landmarks are clearly visible, the person did not move during the exposure, and neither the head nor the feet have been cut off. Include only one person in the photo; keep mirrors, reflections, and other people's silhouettes out of the frame where possible. [7]
If the image is blurred, shows the wrong side, or captures a rotated torso, retake it straight away. Cropping away empty background does not correct perspective, and moving a point manually cannot recover an anatomical detail that was hidden.
You can take two or three photos in case of blur or incidental movement. Do not choose a photo simply because the person looks “straighter” in it.
Use the original image file for upload. Screenshots, photos of a screen, and repeated sharing with heavy compression can make details harder to distinguish. Keep the original before processing the image.
7. Record the setup for the next assessment
A short record is more useful than trying to remember exactly where the tripod stood. You can add the following fields to a photography log or a case note.
| Parameter | What to record |
|---|---|
| Date and context | Date, time, and circumstances relevant to the comparison, such as before or after an exercise session. |
| Device | Phone model, camera used, zoom setting, and phone orientation. |
| Camera height | Distance from the floor to the center of the lens. |
| Distance | Distance from the lens to the chosen line where the person stands. |
| View | Front or right side; label any additional views separately. |
| Stance | Foot and arm positions; habitual or actively corrected posture. |
| Instructions | The wording used before the photo and how gaze direction was specified. |
| Clothing and lighting | Barefoot or wearing specified shoes, clothing worn, and light positions. |
| Analysis | PAX mode, method used, and any changes to the conditions. |
If you are comparing photos taken before and after a session, state that explicitly. When taking follow-up photos several weeks later, reproduce the context as far as possible too.
8. Check the anatomical landmarks in PAX
Take a photo in the app or select a prepared image from the gallery. Then check the crop and the suggested landmark positions. PAX allows manual adjustment of the points; automatically identified landmarks still need to be reviewed by the user. [8]
Zoom in and review each point in turn. Make sure you use the same anatomical landmark at follow-up assessments. If clothing obscures a point or the viewing angle is unsuitable, it is preferable to retake the photo rather than guess its location from the body outline.
After checking the points, save the result as a PDF. Use the same mode, view, and method for subsequent comparisons. In PAX, results from Quick Analysis mode and Full Analysis mode are compared separately. [7]
How to interpret differences between two photos
Differences between results may reflect changes in body position, the way the person resumes their stance, camera setup, and landmark placement. In the study by Dunk and colleagues, participants repeated the photography procedure across several sessions, and individual posture remained variable. The authors emphasize the need to account for baseline repositioning error when interpreting changes. [5]
In the study by Candotti and colleagues involving 24 healthy adults, all studied variables showed high relative reliability, measured by the intraclass correlation coefficient (ICC), across seven frames from one 35-second recording. Most also showed high relative reliability when reassessed on another day. Seven frames from one recording are not equivalent to seven independently repeated standing trials. That result applies to the specific method tested; it does not establish the accuracy of every app. [6]
In practice, a small change in an angle should not automatically be described as an improvement. These studies cannot provide a universal threshold, such as “a difference of 2° is always meaningful.” Even a consistent change in geometry and an improvement in how someone feels are different outcomes and should be assessed separately.
PAX describes geometric measurements on a two-dimensional photograph. The result helps document observations, but it does not, by itself, establish a diagnosis or explain the cause of pain. [7] [8]
Frequently asked questions
Can I take the photos at home?
Yes, if you can secure the phone, provide enough space and light, and reproduce the conditions. If the person fits in the available space only when the camera is tilted substantially, choose another location.
Does the photo need to be square?
Do not crop a photo into a square beforehand if that would cut off the head or feet. PAX does not require a square format for ordinary analysis photos. Check the crop suggested by the app and leave space around the person. [7]
Do I need a grid on the wall?
A grid can provide a visual reference, but it is not required for this checklist. Do not assume that its squares provide a centimeter scale for the body: the wall and the person are at different distances from the lens. Linear measurements in physical units require a separate calibration protocol. [1]
Can I use older photos from the gallery?
You can upload older photos to PAX. However, without a record of the viewing angle, zoom setting, camera height, and instructions, comparison remains limited even after careful landmark placement.
Your first analysis using this protocol
Prepare the space, take front and right-side photos, and record the camera position. Upload the photos to PAX, check the landmarks, and save your first report. At the next assessment, use your log to recreate the same setup.
Scientific and methodological references
[1] Watson AWS (1998). Procedure for the production of high quality photographs suitable for the recording and evaluation of posture. Revista de Fisioterapia da Universidade de São Paulo. 5(1):20–26.
A methodological paper on camera geometry, perspective, and calibration. Its technical recommendations for film cameras have not been carried over to smartphones in this article.
[2] Mota YL, Mochizuki L, Carvalho GA (2011). Influence of the camera resolution and distance in the measures made by the postural assessment software (SAPO). Revista Brasileira de Medicina do Esporte. 17(5).
An experiment using a mannequin. It informs the discussion of camera setup but does not establish the accuracy of photographing people.
[3] Stolinski L et al. (2017). Two-dimensional digital photography for child body posture evaluation: standardized technique, reliable parameters and normative data for age 7–10 years. Scoliosis and Spinal Disorders. 12:38.
A detailed protocol for preparing and positioning children. The age range to which the findings apply is specified in the article.
[4] Ferreira EAG et al. (2010). Postural assessment software (PAS/SAPO): validation and reliabiliy. Clinics. 65(7):675–681.
A study of measurement agreement using existing photographs. The technical measurement error found using an inanimate object cannot be attributed to PAX.
[5] Dunk NM, Lalonde J, Callaghan JP (2005). Implications for the use of postural analysis as a clinical diagnostic tool: reliability of quantifying upright standing spinal postures from photographic images. Journal of Manipulative and Physiological Therapeutics. 28(6):386–392.
A study of variation when people resume a standing position and the resulting limits on interpreting changes.
[6] Candotti CT et al. (2019). Repeatability and Reproducibility of Postural Variables by Photogrammetry. Journal of Manipulative and Physiological Therapeutics. 42(5):372–378.
An assessment of measurement repeatability within a session and between days in 24 healthy adults.
Sources on PAX features
[7] PAX (2026). App features and sample reports. Checked on October 25, 2026.
A public description of analysis modes, manual landmark adjustment, reports, and comparisons. Cropping behavior and compatibility between modes were also checked against the source code for app version 1.10; those developer materials are not publicly available.
[8] PAX (2026). Official PAX listing on RuStore. Checked on October 25, 2026.
Publicly listed features and the app installation page.
Updated