Hou weel dae ye mind an unkent face? Learn a wee set, syne find them amang lookalikes, first in the same view an syne fae anither angle. Aboot 2 minutes, free, nae sign-up. Thir are computer-made faces o fowk that dinna exist.
Bi the Helibrain teamChecked bi Robert(Inglis), PhD candidate in biophysicsPit oot Last checked
Whit dis a face mindin test measure?
A face mindin task asks ye tae learn unkent faces an recognize them later amang ither faces. This gemme separates recognisin the same picture frae recognisin an identity in a new view. Yer result describes this short gemme wi synthetic portraits. It isnae a standardised assessment o ilka-day face recognisin or a clinical test.
Quick facts
- The task: learn 6 faces, ilk yin for 4 seconds, syne mak 12 choices.
- The choices: yin learned identity an unkent foils in ilka group o 3.
- Twa views: the same portrait first, syne anither angle an lichtin o the same identity.
- Yer result: richt choices, accuracy an a brekdoon bi view. Speed disnae add pynties.
- The pictures: original computer-made faces o fowk that dinna exist. Nae camera or photo upload is needed.
Hou the gemme warks
- Lat the portraits load. Ilka image is loaded afore learnin begins. Use a clear screen an yer usual glesses gin ye need them.
- Learn the faces. They appear yin at a time an chyng automatically. Look at ilk body's features.
- Find a familiar face. Tap the face ye learned, or press its numbered key. Ae choice only in ilk group belangs tae the learnin set. Ye can tak yer time chuisin.
- Try the new view. Later choices shaw the same identities fae anither angle. Ye dinna see the learnin pictures alang wi the choices.
- Read yer result. Leavin the tab durin play ends the attempt sae ye can start a fresh yin.
This is a sicht task. Screen readers can announce the controls, but a verbal description cannae keep the face-recognisin challenge. Phone an desktop use the same images an rules; screen size an viewin conditions can still affect performance.
Hou tae read yer result
Accuracy is the nummer o richt choices dividit bi the nummer o choices. The twa view scores shaw whit happened in ilk pairt, wi juist a few answers in ilk yin. Ae mistak can shift the percentage noticeably. There is nae passin mark.
Yince enough eligible players hae feenished this version, the result shaws a below, typical or above comparison band. Till then it shaws the nummer o eligible runs gathered. Comparisons uise ither players' first runs o the same version, wi flagged runs left oot. Website visitors are a self-chuised group, no a population norm. A missin comparison disnae change yer accuracy.
The face set is wee. Playin it again maks the identities mair familiar, even whan their order an roles change. Shared links lat a freend try the same set an order; the freend's displayed score is gien by the link, no checked independently.
Research ahint the task
Duchaine an Nakayama's Cambridge Face Memory Test brocht in a learn-then-recognize format wi unkent faces, includin recognisin them throu changes in the image [1]. Dunn an colleagues' UNSW Face Test brings the challengin face mindin an matchin tasks thegither, showin why a wider set o tasks matters whan assessin individual differences [2].
Our gemme borrows the general idea o learnin an recognisin. It is neither o thae research tests. Its synthetic faces arenae the validated research stimuli; its length, image set an scorin are different. We dinna bring in thae tests' norms or uise this score tae gie an exceptional-ability label or diagnose a condition. The new-view images hae been visually checked for identity consistency, but this bank hasnae been validated independently.
Aksed questions
Is this the Cambridge Face Memory Test?
Naw. This is an original, shorter gemme inspired by the general task format. It uises original synthetic portraits, different trials an comparisons wi fowk that play here. Published Cambridge scores cannae be turned intae this gemme's results.
Are thir photographs o real fowk?
Naw. They are computer-made faces o fowk that dinna exist. We uise nae real photographs or research face sets. Nae names or personal biographies are attached tae them.
Dae a heich score prove an unusual face-recognisin ability?
Naw. It means ye answered mony o thir choices richt. A short gemme, past practice an a wee image bank cannae prove performance across ilka-day situations.
Why micht ma score change?
Attention, weariness, familiarity wi the portraits an screen conditions can change a run. The order o choices changes an aa on a new attempt. Tak tent wi wee differences.
Is the face mindin test free?
Aye. The gemme an the hale result are free, wi nae sign-up or payment.
Related tests
- Emotion kentin testMatch facial expressions tae the labels meant for thaim, fae subtle tae strang.
- Sicht mindin testHou muckle o a pattern o squares ye can mind.
- Wird mindin testInglisHou lang ye can tell new wirds fae anes ye've seen afore.
Want tae wark on this ilka day? The Helibrain app (in Inglis for noo) has a free plan an paid plans.
Cite this page: Face Mindin Test. Helibrain. 2026. https://helibrain.com/sco/face-memory-test. CC BY 4.0.
How this test works
This game measures how well you remember unfamiliar faces. First you learn 6 computer-generated faces, one at a time, for 4 seconds each. Then you see groups of 3 faces and pick the one you learned. The first 6 choices show the same pictures you learned. The next 6 show the same people from another angle and in other light. Your score is the share of right choices. This learn-then-recognize format follows a face memory test published in 2006.
Version 1 · Cite this test:
Helibrain (2026). Face memory test, v1. https://helibrain.com/face-memory-test/methods
Sources
- Duchaine B, Nakayama K (2006). The Cambridge Face Memory Test: results for neurologically intact individuals and an investigation of its validity using inverted face stimuli and prosopagnosic participants. Neuropsychologia 44(4):576–585. pubmed.ncbi.nlm.nih.gov/16169565/
- Dunn JD, Summersby S, Towler A, Davis JP, White D (2020). UNSW Face Test: A screening tool for super-recognizers. PLOS ONE 15(11):e0241747. journals.plos.org/plosone/article?id=10.1371/journal.pone.0241747
How we made this page: scrieved bi the Helibrain team fae the studies listed abuin, an checked bi Robert(Inglis) (PhD candidate in biophysics) on 11/4/2026. We update it when the soorces chynges.