Orientation & setup
- State what the toolkit is for, and what it is not
- Activate a licence and move through the scan workflow
- Use the desktop shortcuts
What the toolkit is
The Iridology Analysis Toolkit captures images of both eyes and measures pupil and iris features: pupil-to-iris ratio, pupil shape and position, collarette (ANW) position, and local deformations of the pupil margin. It overlays historical iridology charts and produces PDF, text, JSON and Excel records.
It is a research and educational tool for trained practitioners. It is not a medical device, and no regulator (FDA, CE, TGA or equivalent) has cleared it for diagnosis. Every number it produces is an observation you interpret. You make the clinical decisions.
Licensing
| Tier | Duration | Machines |
|---|---|---|
| Trial | 14 days or 20 analyses, full features | 1 |
| Standard | 1 year | 1 |
| Professional | Lifetime | 1 |
| Enterprise | Lifetime | Multiple |
On Windows, paste your key into License Key and click Activate. Activation needs an internet connection once, to bind the key to the machine. After that the app runs offline.
The scan workflow
The patient form asks for name, age and sex. Main complaints and the clinic name are optional. Age matters more than it looks: the age-normalised pupil comparison needs it. The clinic name is remembered between sessions.
Desktop shortcuts
| Key | Action |
|---|---|
| F11 | Toggle fullscreen |
| Ctrl + H | Open scan history |
| Esc | Go back or close a dialog |
The settings icon in the title bar is available on every screen. It holds the therapy-module switches, the clinic name, language (20 languages) and PDF options.
Capturing great images
- Choose the right camera source for the situation
- Read the quality gate's feedback
- Frame and light an eye so the measurements hold up
- Import phone or archive photos correctly
Almost every questionable result traces back to the image. Two minutes spent on the capture saves you from interpreting a measurement of glare, eyelash or the wrong circle.
Capture order
The toolkit captures the right eye (OD) first, then the left eye (OS), which is the usual clinical notation. Analysis needs both eyes.
Camera sources
| Source | Use it when |
|---|---|
| Quality-gated rear / front camera | Phone or tablet. This is the recommended mobile mode, because it rejects bad frames automatically. |
| Manual camera | You need to control the shutter yourself, you're using a macro attachment, or unusual lighting makes the gate reject good frames |
| USB / UVC iriscope | Dino-Lite or a compatible iriscope, on Windows or on Android over USB OTG |
| Import from gallery | Re-analysing archived photos, or images taken elsewhere |
The quality gate
In quality-gated mode every frame is checked against five criteria at once. When all of them pass, the border turns green and the photo is taken automatically.
| Criterion | Must be | Catches |
|---|---|---|
| Sharpness | ≥ 100 | Motion blur, missed focus |
| Brightness | 30–230 of 255 | Under- or over-exposure |
| Contrast | ≥ 30 | Flat, detail-less frames |
| Pupil confidence | ≥ 30% | No detectable pupil (blink, off-target) |
| Centre offset | ≤ 25% of frame | Eye not centred |
RED at least one criterion is failing, so read the message. AMBER borderline, so make a small adjustment. GREEN everything passes and the photo is taken.
Using an iriscope
- On Windows, install the DNVideoX driver before you connect a Dino-Lite for the first time.
- Work at about 2–3 cm from the eye (Dino-Lite AM4115ZT).
- Use the built-in LED ring. Extra lamps add corneal reflections.
- Adjust the zoom slider until the pupil margin is sharp. The margin is what the deformation and P/I measurements depend on.
Capturing with a phone
Hold the phone upright (portrait) and let the eye fill the frame. A wide landscape strip leaves the iris detector very little room to search: in testing, detection confidence fell from about 0.81 in portrait to 0 in a very wide crop of the same eye.
Importing photos
Imports open a 4:3 cropper. Crop tightly around the eye so the iris takes up a good share of the frame. An arm's-length phone photo with a small iris in the middle cannot be rescued by enlarging it, because the detector looks for the iris edge at a set fraction of frame width. Cropping changes that fraction. Scaling does not. The app copies the imported file into its own folder, so your original is never changed.
Do
- Ask the patient to fix their gaze on a distant point
- Brace your hand or the device
- Shift a little to move corneal reflections off the pupil margin
- Let the iris fill at least half the frame width
- Use the same device and distance at every visit
Avoid
- Overhead lights reflecting in the cornea
- Pre-cropping phone photos into wide strips
- Switching to manual mode just to get past a gate that is right
- Mixing devices or framings in one patient's timeline
Reading the measurements
- Interpret P/I ratio, ellipseness, decentration and anisocoria against their bands
- Grade FLAT and PROT zone findings
- Read the ANW ratio and the hybrid confidence score
- Pupil
- Dark centre. The teal line is its diameter.
- Collarette
- The ANW, drawn in amber here as it is in the app's chart overlay
- Limbus
- Outer edge of the iris. The bottom line is the iris diameter.
A schematic, not to scale. Every ratio below divides one of these diameters by the iris diameter, so the numbers don't depend on image size.
P/I ratio (pupil ÷ iris × 100)
| P/I | Label |
|---|---|
| < 15% | Miosis (very constricted) |
| 15–19.9% | Constricted |
| 20–29.9% | Normal (adult, indoor light) |
| 30–39.9% | Dilated |
| ≥ 40% | Mydriasis (very dilated) |
Pupils get smaller with age, so the results also compare against an age band (adults 18–39 expect about 4.5–5.5 mm, 60 and over about 3.5–4.5 mm). That comparison only appears if age was entered.
Try it: P/I and anisocoria
Enter diameters in pixels for each eye. The example values are made up. Edit them to see how the bands change.
Zone findings: FLAT and PROT
The pupil margin is checked zone by zone for places where it departs from a circle. FLAT means the margin is pulled inward at that zone. PROT means it bulges outward.
| Deviation | Severity |
|---|---|
| < 1.5% | Not reported |
| 1.5–2.9% | Within limits |
| 3.0–5.9% | Mild |
| 6.0–9.9% | Moderate |
| ≥ 10% | Significant |
ANW (collarette) ratio
| ANW ratio | Status | Traditional reading |
|---|---|---|
| < 25% | Spastic | Ring drawn inward, sympathetic dominance |
| 25–35% | Normal | Balanced tone |
| > 35% | Atonic | Ring expanded, parasympathetic dominance |
The difference between the two eyes' ANW ratios is normal up to about 5%. When one eye reads Spastic and the other Atonic, the app flags a Functional Frustration pattern.
Pupil shape, position and symmetry
| Metric | Flag when | Notes |
|---|---|---|
| Ellipseness (minor ÷ major axis) | < 95% | The form is reported as horizontal, vertical or diagonal oval, chord-like or irregular |
| Decentration (offset ÷ iris radius) | ≥ 5% | Direction (nasal, temporal, frontal, basal…) and angle in degrees clockwise from 12 o'clock |
| Anisocoria (|OD − OS| P/I) | ≥ 2% mild · ≥ 4% moderate · ≥ 8% severe | From moderate upward, the app raises a TBI reference flag. It is a research reference, not a diagnosis. |
Hybrid confidence
Each eye goes through two independent pipelines: classical computer vision and an ML model. The confidence score combines capture quality, how well the classical pipeline found the boundaries, and how closely the two agree. Above 80%: the pipelines agree and the image is good. Below 50%: treat the result with caution and recapture if you can.
The ANW Velchover wizard
- Explain why local ANW patterns need your confirmation
- Work through the six screens, or skip them for the right reasons
After analysis, and before the results appear, the wizard shows you the local collarette patterns the app detected in each sector. Detection on its own is noisy: 88.9% of automated geometric detections were false positives. That is why the report uses only the patterns you confirm.
The six screens
| Screen | Pattern category | Clock range |
|---|---|---|
| 1 | Drawing-out, right AND left | 10:30–1:30 |
| 2 | Drawing-out, right iris only | 10:30–1:30 |
| 3 | Drawing-out, left iris only | 10:30–1:30 |
| 4 | Drawing-out, right OR left | Several ranges |
| 5 | Drawing-in, right OR left | Several ranges |
| 6 | Flattening, right AND left | 10:30–1:30 |
Each screen shows both eyes with radial guide lines and a neutral circle. The collarette the app fitted is deliberately left off these screens so it doesn't steer your judgement. Chips come pre-ticked from the detector; tap to change them. Rules stop you from selecting patterns that contradict each other.
When to skip
On the first screen, Skip records an empty confirmation list, so no Velchover patterns go into the report. Skip when glare or poor focus makes the collarette impossible to judge. If one eye's collarette wasn't detected at all, the wizard doesn't appear.
Do
- Look at the image yourself before you accept a pre-ticked chip
- Keep screens 1 and 6 separate in your mind: same clock range, opposite deviations
- Skip rather than guess on a poor image
Avoid
- Clicking Next through every screen without looking
- Wording findings as "indicates…". The app and report say "historically linked to…".
Visual tools & chart overlays
- Choose the right visual tool for a question
- Fit a historical chart to a patient's eye and add structured findings
Image tools on the results screen
| Tool | Use it to |
|---|---|
| Enhance (CLAHE) | Bring out crypts, lacunae and fibre texture without blowing out bright areas |
| 3D Relief | See bright raised fibres as ridges and crypts as valleys. Drag to rotate, scroll to zoom. |
| Heatmap | See where pigment is concentrated: heterochromia sectors, a darkened ciliary zone |
| Annotate | Draw on the iris. Undo, Clear, and Save PNG (saved as Annotated_<Name>_<date>). |
| Visual Compare | View OD and OS side by side. Mirror OD turns both nasal sides inward. Sync Pan links zoom and pan. |
| Gabor / LBP | Texture maps for fibre orientation and stroma pattern (research use) |
| Research tab | Unwrapping, crypt and furrow detection, SSIM, GLCM and more. Every detection there is a suggestion for you to verify. |
Historical chart overlays
Eight authors are available: Velchover, Angerer, Bourdiol, Gunther, Jausas, Roberts, Jensen and Vida-Deck. Organ names appear in the app language. Open Chart shows one eye; Bilateral Chart shows both under the same chart.
| Control | What it does |
|---|---|
| Auto-fit | Sizes the chart to the detected iris automatically |
| Scale slider (0.3×–2.0×) | Fine-tune the chart size |
| Opacity slider | Fade the chart so you can see the iris underneath |
| ANW ring slider | Move the amber collarette ring onto the collarette you can see |
| Warp to eye / chart | Warp to eye (the default) bends the chart's pupil, collarette and limbus rings onto this patient's anatomy, so the drawn borders match the zone under the cursor. The toggle resets to warp to eye each time the dialog opens. |
With warp to eye on, nothing is drawn inside the patient's pupil. That is intentional, because there is no zone there.
Recording a finding with the Anomaly Picker
Tap a zone → open Add Finding → choose the anomaly type and, if there is one, a subtype → Add to Notes. The finding goes into Observer Notes as [OD] [Zone] Type › Subtype and then into the PDF and text reports.
Reports, history & therapy panels
- Find, reopen and export past scans
- Set up the PDF and the optional therapy modules
- Add a constitutional type
Scan history
Every analysis is saved automatically when the results screen loads, so there's no save button. Open history with Ctrl+H or the Scan History button. From there you can reopen a record, regenerate its PDF, export the whole history to Excel, or open the Serial Timeline, which charts OD/OS P/I, ellipseness and confidence across one patient's visits.
Exports
| Format | Best for |
|---|---|
| The full report: photos, metrics, zones, ANW, notes, constitution, therapy, disclaimers | |
| TXT | Pasting into an EHR or an email. No photos or therapy sections. |
| JSON | Research pipelines and databases |
| Excel | The whole scan history, one row per scan |
PDF settings: include images (on by default), clinic name banner, report language (follows the app language) and auto-save PDF (off by default).
Therapy panels
The herbal, nutrition (7-colour diet), chiropractic (spinal segments) and TCM (Five Elements) modules are all off by default. Switch them on in Settings. A panel appears only when a zone finding of Mild or higher points to an organ system, and the most severe zones are listed first.
Constitutional type (optional)
On the Both Eyes Captured screen, scroll down to the amber "Constitutional Type (Optional)" panel and pick one of the 34 types, grouped as Lymphatic, Hematogenic, Mixed, Connective Tissue, Neurogenic and Cardio-Pathological. A constitutional panel then appears in the results and the PDF. A constitution describes a tendency, not a disease, so only choose one if you have constitutional training.
Best habits, tips & troubleshooting
A routine for every scan
- Fill in age as well as name. Without it you lose the age-normalised comparison.
- Set up the room: dim the overhead lights and ask the patient to fix on a distant point.
- Capture OD, then OS, the same way: same device, distance and zoom.
- Look at the confidence score before anything else. Below 50%, retake.
- Go through the ANW wizard properly, or skip it on purpose.
- Fit a chart, add findings with the Anomaly Picker, and check the Observer Notes.
- Export the PDF. The history record is already saved.
Troubleshooting
| Symptom | Likely cause | Fix |
|---|---|---|
| Gate never turns green | Glare, blur or an off-centre eye | Read the red message, change the angle, brace your hand. Use manual mode only if the lighting is genuinely unusual. |
| Imported photo fails, or the P/I looks odd | The iris is too small in the frame, or the image is a wide strip | Crop tightly in the 4:3 cropper. Enlarging the image doesn't help. |
| P/I jumps between visits | Different framing or device | Standardise the capture before you read it as a real change |
| ANW wizard didn't appear | One eye's collarette wasn't detected | Recapture with less glare near the pupil margin |
| Chart borders don't sit on the anatomy | The ring or scale is off, or warp is set to chart | Switch to warp to eye, align the ANW ring, then adjust scale |
| Therapy tab missing | The module is off, or no finding reached Mild | Check Settings. No panel is the right result when nothing qualifies. |
Tips from the field
- Question findings that disagree with what you see. If the app reports something your own reading doesn't support, look at the capture and the fitted rings before you believe either one.
- Treat the timeline as serious data. It's only as reliable as your consistency from visit to visit, so keep the same device, distance and zoom for each patient.
- Use Mirror OD with Sync Pan in Visual Compare to scan both eyes for symmetry quickly.
- Annotate before you export. A saved PNG with the lacuna circled explains a finding faster than a paragraph.
- Enhance, then Heatmap: CLAHE shows the structure, and the heatmap shows the pigment on top of it.
Final exam
Twenty multiple-choice questions covering all seven modules. Answer every question, then submit to see your score and an explanation for each answer. You can retake it as often as you like, and your best score is saved in this browser.
This course teaches how to operate the Iridology Analysis Toolkit. The traditional associations it mentions are historical interpretations, not diagnostic claims. The toolkit is a research and educational tool, not a medical device.