No—the Face Memory Test uses 15 original geometric SVG illustrations, not photographs, and it does not identify real people. If you have seen this described as a “face sequence game,” the more precise description is a five-round visual-recall challenge. Each round shows one fictional profile for two visible, one-second countdown ticks, or lets you choose Recall now to hide it sooner. Once the target disappears, you pick the same drawing from three alternatives. The game does not request camera access, accept a portrait file, compare an upload, search an identity record, or run biometric matching. Its shapes are authored puzzle features, not measurements of a human face. The named profiles Alder through Opal are fictional fixture identifiers: they do not stand for a person, demographic group, diagnosis, emotion, or biometric template. The task is simply to decide which original illustration was displayed a moment earlier. It does not recognize a face, estimate identity, infer sensitive traits, or claim to measure memory ability. Everything in the game runs locally, so “fictional face” here means a code-drawn visual profile rather than a photograph of a made-up or real person.
What the game actually displays
The visual set is fixed and code-drawn. Face Memory Test assembles 15 abstract profile specifications from a small authored vocabulary instead of selecting pictures from a portrait library. Each fictional profile combines these puzzle elements:
- Head outlines: round, oval, or square.
- Hair: five authored shapes.
- Brows: three shapes.
- Eyes: three shapes.
- Noses: three shapes.
- Mouths: four shapes.
- Optional details: cheek marks or freckles.
These are immutable SVG specifications, not parameters sampled from a person. A hair shape is one of the game’s visual cues, while optional freckles are marks a player can compare; neither is evidence about someone’s identity or biology. All 15 drawings are original geometric illustrations. No photograph, portrait file, stock image, generated-person asset, or external visual is loaded.
Each SVG has a stable view box that reserves its dimensions and avoids layout shifts. Semantic foreground, surface, accent, and border variables also keep the artwork visible with the site’s themes. These implementation details explain how the profiles appear directly in the game without downloading an image file. They do not turn the illustrations into portraits.
How it differs from face identification
The crucial distinction is between recalling a seen design and recognizing an unknown person. Here, the target already belongs to the current round. Your job is to select that exact fictional profile from the three-button tray after it disappears. Nothing is compared with a live camera image, an uploaded portrait, a facial template, or an identity database.
| Aspect | Face Memory Test | Excluded from the game |
|---|---|---|
| Visual source | 15 original geometric SVG profiles | Photographs, portrait files, stock images, or generated-person assets |
| Player task | Recall the illustration displayed a moment earlier | Recognize, identify, or verify a real face |
| Identity data | No identity record or biometric template | Upload comparison, identity lookup, or biometric matching |
| Output | Round progress, misses, and a fixed score | An identity, demographic, emotion, diagnosis, or sensitive-attribute label |
| Processing | Local deterministic game logic | A camera model, account, API, machine-learning model, or network call |
This also explains why the phrase “face sequence game” can suggest the wrong idea. A sequence of five memory rounds does not make the profiles a facial-analysis sequence. They are fixed puzzle records with fictional labels and code-drawn features.
How to complete the five-profile challenge
Start the run and use the same visual-matching process in every round:
- Study the target during Reveal. One fictional profile appears, and no answer buttons are visible. Check the combination of head shape, hair, brows, eyes, nose, mouth, and any optional marks. The target remains through two visible, one-second ticks. Select Recall now when ready; pressing Enter or Space during Reveal performs the same transition.
- Wait for the target to unmount. The game changes to Recall and mounts three distinct profile options. Decide which drawing exactly matches the one you just studied. The question concerns an authored illustration, not the identity of a person.
- Choose the matching profile. Select its large illustrated button or press its corresponding number from 1 through 3. Keyboard users can also move the visible selection with Left and Right, then confirm with Enter or Space. Up and Down deliberately do nothing because the options occupy one row.
- Continue according to the miss rules. A correct choice adds 200 points and opens the next target. One distinct wrong profile can be recovered, but choosing that same wrong profile again does not create another error. A second different wrong profile closes the run and disables further answers.
- Match all five targets. Complete every round before a second distinct miss to finish with exactly 1,000 points. If you want to restart, the shared Restart control returns the game to the first target and restores its two-tick countdown.
Reveal and Recall controls compared
The two game states have deliberately separate controls. Knowing which state is active helps explain why answer keys do nothing during Reveal and why the original drawing is unavailable after Recall begins.
| State or control | Available action | Defined result |
|---|---|---|
| Reveal | Study the single target or choose Recall now | The target unmounts after two ticks or the early transition |
| Enter or Space during Reveal | Hide the target early | Recall opens with three answer buttons |
| Recall | Use a large option button or keys 1 through 3 | The selected current-round profile is scored |
| Arrow selection | Move Left or Right, then confirm with Enter or Space | The visible selection is confirmed; Up and Down have no effect |
| Restart | Return to the opening target | The two-tick Reveal countdown is restored |
| Escape shortcut | Double-press Escape | A simulated spreadsheet boss screen opens and can be resumed the same way |
Reveal and Recall are isolated states. During active Reveal, the target is not concealed behind opacity, clipping, alternative text, a data attribute, or a hidden option list. Unknown identifiers, options from another round, profiles outside the current tray, and actions attempted in the wrong phase are atomic no-ops. Moving the pointer outside the three-button tray does not create a selection.
The controls also provide textual option labels and visible keyboard focus. Status text announces the round, number of matches, and errors without relying on color alone. These accessibility features support the visual-recall task, but they do not convert it into an ability assessment.
Scoring, errors, and fixed answer order
Every clean match adds exactly 200 points. The full scoring calculation is straightforward: 5 correct matches × 200 points = 1,000 points. Reveal duration does not create a timer bonus, and there is no hidden score adjustment.
The five rounds are deterministic. Answer positions are not shuffled, and no random portrait is inserted. For audit purposes, the correct positions among the labeled buttons shown to the player are 2, 1, 3, 2, and 1. There is also no adaptive model that changes a profile or answer based on earlier choices.
Wrong attempts are recorded by round and profile. After one distinct mistake, choosing the same wrong profile again in that round is deduplicated, so the error count does not rise. Selecting a second different wrong profile triggers the terminal state and disables further answers. A correct selection remains available during the recoverable first-miss state.
Local processing and browser storage
The game makes no camera request and provides no file-upload control. It also uses no account, API, machine-learning model, cookie-based identity, or network call. The game logic processes the fictional profiles locally, so there is no submitted photograph or identity record for the tool to recognize.
The only information that can remain in the browser is the shared best completed score. The illustrations use themed SVG variables rather than downloaded image assets, while their stable dimensions prevent the target from shifting the page as rounds change. These details support local, repeatable play without turning the game into a biometric or portrait-matching service.
What the final score means
A completed 1,000 points means that all five fixed fictional targets were matched from their current three-option trays before a second distinct miss occurred. It records success under the game’s deterministic rules. It is not a norm-referenced memory score, a diagnosis, or a claim that one person remembers better than another.
For the original search question, the distinction is definitive: Face Memory Test does not use photographs and does not identify real people. It is a visual-recall game built from 15 original geometric SVG profiles, with fictional names serving only as fixture identifiers.