IAI Forensic Art common mistakes
IAI Forensic Art · Understand the traps and how to avoid them.
Treating the facial reconstruction as a portrait and adding features or expressions that are not supported by the skull, such as smiling mouths, defined eyelid creases, or specific hairstyles.
Candidates come from illustration or portrait backgrounds and instinctively 'finish' a face, forgetting that the reconstruction must reflect only what the osteology supports and that soft-tissue features are probabilistic.
What to do instead
Before adding any feature, state the skeletal evidence for it. If there is no skeletal anchor (e.g., for eye shape, lip thickness, or hair), leave it neutral or explicitly labelled as speculative. Practise writing a short 'evidence justification' for every feature you render.
Using average tissue-depth tables as if they were exact per-individual measurements, placing a single depth value at each landmark without accounting for age, sex, ancestry, or body mass variation.
Depth tables are convenient and memorisable, so candidates treat them as deterministic constants rather than population means with wide ranges.
What to do instead
Learn which variables (sex, ancestry, age, emaciation) shift depth values, and practise applying a range rather than a point value. Be able to explain in writing why a given depth was chosen for a given landmark.
Misplacing the anatomical landmarks themselves, e.g., putting the pupil centre too high or low in the orbit, or locating the mouth at the wrong vertical position relative to the teeth.
Candidates memorise rules of thumb (such as pupil at mid-orbit) without understanding orbital asymmetry and variation, and they skip checking landmark placement against the actual cast or scan.
What to do instead
Drill landmark location on multiple skulls of differing ancestry and age, and cross-check each landmark against at least two independent cues (e.g., orbital margin and dental occlusion for mouth position). Document the rationale for each placement.
Presenting an age progression as a single definitive image rather than a range of plausible outcomes, and failing to anchor changes to the reference photograph's specific features.
Candidates want a polished result and produce one 'best' face, overlooking that ageing is nonlinear and highly variable, and that the family and investigators need to understand uncertainty.
What to do instead
Produce multiple age-progressed variants showing plausible weight, hairline, and lifestyle changes, and keep the original identifying features (e.g., ear shape, dental anomalies) consistent. Write a short limitations statement for each case.
Confusing the goals of post-mortem identification imagery with those of a facial reconstruction or age progression, e.g., 'beautifying' a decedent's image or altering injury-related features that are needed for recognition.
Candidates apply cosmetic retouching habits and do not distinguish between restoring a presentable image and preserving identifying detail; the emotional weight of the task also pushes toward idealisation.
What to do instead
Study the distinct purposes of post-mortem depiction versus reconstruction and list which features must never be altered (e.g., scars, tattoos, dental work, asymmetries). Practise restoring image quality (lighting, colour balance) without changing anatomy.
Ignoring craniofacial asymmetry and building a symmetrical reconstruction, then being unable to explain discrepancies between the reconstruction and the skull.
Symmetry is visually pleasing and easier to draw, and candidates assume the skull is essentially symmetric when in fact most skulls show measurable asymmetry.
What to do instead
Measure and record asymmetry on the actual skull or cast (e.g., compare left and right orbital and zygomatic dimensions) and deliberately reproduce it. Include a symmetry check as a standard step in your workflow.
Skipping or superficially documenting the case record, so that the final image cannot be traced back to the measurements, references, and decisions that produced it.
Candidates focus on the final artwork and treat documentation as bureaucracy, not realising that a forensic image must be defensible and reproducible for court or peer review.
What to do instead
Adopt a fixed documentation template (landmarks, depths, references, software settings, versions, date, examiner) and complete it as you work, not afterwards. Practise reconstructing a case from someone else's record to see what is missing.
Using unsuitable or unverified reference photographs (different angle, lens distortion, heavy makeup, or a different individual) as the basis for a comparison or progression.
Candidates grab the most available image and underestimate how pose, focal length, and expression distort facial proportions, leading to comparisons that are not like-for-like.
What to do instead
Establish criteria for acceptable reference images (neutral expression, comparable pose and lighting, known date) and reject or correct those that fail. Learn to normalise for angle and lens effects before comparing.
Overstating the certainty of a conclusion in the written report, e.g., asserting a positive identification or a definite match from a reconstruction or superimposition.
Candidates conflate 'consistent with' and 'identified as', and they fear that hedged language will look weak or unhelpful to investigators.
What to do instead
Learn and use the standard scale of conclusions (e.g., consistent with, cannot exclude, excluded) and match each statement to the strength of the underlying evidence. Have a colleague challenge any sentence that implies certainty beyond the data.
Neglecting the practical constraints of the medium and workflow, such as producing an image that cannot be reproduced at the required scale, or failing to keep original files and version history intact.
Candidates trained in illustration work in whatever format looks best on screen, and they overwrite files without considering chain-of-custody and reproducibility requirements.
What to do instead
Work to defined output specifications (resolution, scale, file formats) and maintain immutable originals plus a clear version log. Treat every file as potential evidence and rehearse producing a print-ready, reproducible final image from the archived source.