A werewolf needs somewhere for the moon to appear. A background scare needs a place to hide. A person stepping into a mech suit needs room for the suit and a way to leave the scene.
Those decisions depend on your photograph. MagicLens Visual Intelligence is designed to read the scene before rendering, looking for the places and poses that can carry the effect. The aim is a Momento that feels made for the picture you supplied.
A moon belongs in the sky
In this street portrait, there is open sky above the buildings, but branches fill one corner. The check chooses a clear patch for the moon and keeps it away from the person's head. It leaves the blocked corner alone.
An indoor photo asks for a different decision. Sky visible through a window does not turn the room into an outdoor setting. In our café example, the Werewolf scene kept the café's own light rather than adding a moon indoors.
The setting should give the transformation a place to happen, not bend to make room for it.
Sometimes there are more places NOT to hide than to hide
A background scare needs one good hiding place, and a backyard at night offers more than a dozen candidates. Visual Intelligence reads them all: the string lights, the fence, the fire pit, the foliage, the windows, the ground and the shadows.
Most of them will not work. A figure cannot stand in front of the person, climb out of a window that may be closed, or hide behind shrubs with no visible ground to stand on. So the check rules places out until a single spot is left: the dark corner where the fence meets the house, behind the person, with ground underfoot and the house corner for cover.
That corner is where the scare comes from.
Sometimes the useful answer is another photo
And sometimes it just won't work, so Visual Intelligence rejects the photo. A close portrait in a garage may have nowhere behind the person to hide a figure. A crowded living room may offer neither clear floor space nor a usable exit for a mech.
We want that explanation to be useful. An answer such as “there is nowhere behind you for this effect” gives someone a reason to choose another picture. It also lets us change the plan before making a film that the scene cannot carry.
A closed door can still be an exit
A photo describes one instant. A film can change what happens next.
That is why a closed door can work for the Mech Suit: the scene can include opening it. The check needs to recognise the door, its position and the space around the person, then choose an exit that fits the action.
In a separate empty-park test, the check selected an upward launch. The resulting film was approved in our development review. Different surroundings led to a different way to tell the same kind of story.
Measuring the space between people
A Mech Suit needs room to take off, so Visual Intelligence measures the distance between the person who gets the suit and everyone else in the frame. A take-off fires its flight burners, so the plan treats anyone within a head's width of the wearer as too close; someone a few steps behind is clear.
In this test, the woman in the red jacket is the one who gets the suit. The man beside her measures less than a head's width away in the picture, so the plan proposes removing him before the effect begins, and asks the user first. The man behind her is far enough back to stay.
The same measuring applies to the person themselves. Arms pressed against the body leave too little room for a suit to form, so one test asked for the arms to hang a hand's width from the sides while the feet stay in place.
What we are still improving
Reading a flat photograph has limits. One test treated a porch roof as an obstacle above the people even though it was behind them. Another needed a sharper instruction before it noticed arms held close to the body.
The plans give us something concrete to inspect and correct. We can see which spot was selected, why another was rejected and where the reading of the scene went wrong. The overlays in this article are development evidence; they do not imply that every planning view is available as a customer control.
The people-removal and pose-adjustment examples here show planning decisions. The finished effect still has to follow that plan: noticing a problem and solving it in a film are separate steps.
If you are curious about what the planning serves, see the Halloween Momentos, read how we built the Werewolf transformation, or start with What is a Momento?. Every Momento begins with the same question: what can this photograph become?