Concept

reconstruction ambiguity — where it appears

What a set of pictures leaves undetermined — a scale, a similarity, or seven numbers — however many of them there are. Which of them it is depends on what was assumed rather than on how many pictures there were, and adding views closes some of it and never the reflection.

Named by 34 essays across 11 fields — each of them below, with the objects they name alongside it.

× 4.6e+6seven flatand the rest stiffsingular value ÷ the largest, log scale, smallest firstσ₈/σ₇ = 4.6e+6168 parameters · 528 residuals

Seven numbers no picture can name

Shift a whole reconstruction by a metre and a half, turn it half a radian, scale it by 2.7, and every photograph of it stays where it was to a hundredth of a billionth of a pixel. Move one point by fifty millimetres and they move by two thirds of a pixel.

manyviews · Gauge
near edge 155 pxfar edge 196 pxrecovered: ratio 1.000000, corners 1e-14° from squarecorrect from 12 cm, at 160 mm wideplane tilted -45°

An inverse perspective is a leaning plane

Ask a divergent construction what solid it depicts and it answers: a rectangle, four right angles, near edge equal to far. What the splay encodes is not the shape but the plane's tilt — and a real square on a plane leaning toward the camera really does photograph with its far edge wider.

choices · Divergence
010203040-2502550disparity on the screen — millimetreswhere the point is depicted — metres from the eyesd = 63 mm — the eyes' separationon the glassscreen at 2.0 m · eyes 63 mm apartthe ceiling is the head, and it does not move when the screen does

Two pictures on one screen

A stereoscopic display puts a point where two sightlines cross, so the depicted depth is b·D/(b−d) and the disparity that reaches infinity is exactly the separation of the reader's eyes — 63 mm, at any screen distance whatever. The depth budget is set by the width of a head and by nothing about the scene.

screen · Stereodisplay
points in front of both camerasR₁, t0 of 44180° from the truthR₁, t reversed0 of 44180° from the truthR₂, t44 of 44chosenR₂, t reversed0 of 44180° from the truththe chosen pose is the true one to 0.0e+0°cheirality, not algebra, picks it

Four cameras fit, and one of them can see

The essential matrix does not determine a camera pair. It determines four, all of which reproject every correspondence exactly, and the thing that picks one is not more algebra — it is the assumption that the photographer could see what was photographed.

twoviews · Cheirality
-10-50-2-10baseline (m, log scale)worst deviation in the recovered shape (log scale)exact pointsread to 1 pxat 2 mm: 4.7e-9 exact, 2.9e+1 read to 1 px6e+9× apart

A turn of the head is not a step sideways

The textbook says a short baseline makes reconstruction ill-conditioned. Measured, the algebra does not notice — a two-millimetre baseline recovers the courtyard to nine digits from exact marks. What fails is the ratio of baseline to reading error, and it fails without refusing.

depth · Baseline
as recovered0.308 across× 3.4 — same two pictures1.046 acrossworst ratio deviation 6.3e-14size fixed only by a supplied 0.80 m

Two views give shape and no size

Every pairwise distance ratio in a courtyard recovered from two photographs matches the real one to fourteen digits. The courtyard's actual size is not merely uncertain — it is absent, and a reconstruction three and a half times larger fits the same two pictures exactly as well.

twoviews · Scale
the shadow lines meet below the horizon — a lamp in the roomhorizoncorrect from 25 cm, at 160 mm wide3 posts · foot 351 px below the horizon

The lamp, out of the picture

Two posts and their shadows put the light's position in a photograph, exactly, with nothing given but the drawn lines and the camera's own horizon. Two posts. One gives a residual of zero and an answer that is a whole one-parameter family — the sharpest counter-example there is to the idea that a small residual means a right answer, met again in a new field.

light · Lightrecovery
left pictureright pictureevery one of them fits every mark1.6e-6 px

A flat scene fixes no second eye

Eight marks on one plane leave the eight-point design matrix two ranks short, so a two-parameter family of fundamental matrices satisfies every mark exactly — three of its members are 1.41 apart after normalisation and all three fit to 1.6 × 10⁻⁶ pixels. The same photographs determine the plane's own map from four marks, to 5.9 × 10⁻¹³ pixels.

twoviews · Planar
481632641282560.3131030100how far away the scene is (m, log scale)pixels of parallax, and degrees of pose error (log scale)parallax, pxpose error, degreesparallax goes as distance to the -0.977 distances

Far enough away, a pair is one eye

Hold the baseline and walk the scene away, and the parallax a single homography cannot explain falls as the distance to the power −0.968 — one over the distance, which says the ratio of baseline to depth is the whole of it. The recovered translation direction follows it down, from 3.3° at four metres to 74.5° at two hundred and fifty-six.

manyviews · Degenerate
the needle: along the rays' bisectorreprojection minimum, both framesmidpoint, world framemidpoint, changed framemidpoint moved 0.203 mm · minimum moved 1.0e-15 mtruth 15.5 mm along the needle

The midpoint is a choice of ruler

Two photographs do not change when the world is measured with a different ruler, so an answer that belongs to the photographs cannot change either. The midpoint of two skew rays does: a threefold stretch moves it 0.203 mm and a projective frame 1.503 mm, while the point that minimises reprojection error stays put to 10⁻¹⁵ m. Both are 15.5 mm from the truth, which is the part a choice of route does not touch.

depth · Triangulation
what was drawnthe solid it depicts — reading 1 of 2xyzlooked at 0.00° off the normalcube edge 1.0000 of the drawn unitresidual 1e-16

The drawing does not say which corner is nearer

The Necker cube is filed under optical illusion, as though the flipping were something the eye does. It is not: a parallel drawing of a cube is a drawing of exactly two cubes, mirror images of each other, and they project to the identical picture to the last bit. Perspective rules the second one out at a rate exactly inverse in the eye's distance, and never entirely.

parallel · Reversal
-2e-7-1e-70-0.800-0.700-0.600-0.500α, the mix of the two nullspace directionsthe determinant that a fundamental matrix must make zeroa cubic with three real roots3 matrices, all exact

Seven marks, three answers

Seven correspondences leave a two-dimensional nullspace, and the requirement that a fundamental matrix be singular is a cubic in the mix — one or three real roots. Here it has three, and all three satisfy every one of the seven marks to 8.9 × 10⁻⁹ pixels. The eighth mark, withheld, separates them by more than an order of magnitude.

twoviews · Eightpoint
floor: tilt 36.7°footstool: tilt 65.1°table: tilt 54.2°book: tilt 78.2°assumed camera 520 pxspread 41.4°

Four surfaces, and no one camera that draws them

Read under one assumed camera, the floor, footstool, table and book of a constructed divergent picture imply tilts of 36.7°, 65.1°, 54.2° and 78.2°, where one camera photographing four parallel surfaces gives each of them 35.0°. But the spread between the tilts is 46.8° under a 260 px lens and 6.7° under a 5,000 px one, so it measures the lens as much as the picture. The measure that owes nothing to a lens is on the page: the nearest drawing one camera could make moves the far corners by 17.6 px.

choices · Divergence
both explain both photographs exactly11.4 m apart, median

The surface two pictures cannot separate

There is a quadric through both camera centres on which two genuinely different motions draw identical pictures. Built explicitly, forty-two marks satisfy both epipolar geometries to 2 × 10⁻¹³ pixels, and the two scenes they reconstruct place the same mark at 15.6 metres and 35.1. The design matrix's nullspace has two dimensions rather than one, which is the seven-point situation arrived at from the other side.

twoviews · Critical
0240.50011.50tolerance allowed in the redraw (px, log scale)fewest cameras the picture needs3 at 4.63 px2 at 11.11 px1 at 17.63 px4 surfacesone camera at 17.63 px

A camera count needs a tolerance

Asked how few cameras could have drawn a splayed picture, a constructed stack answers four, three, two or one depending on how many pixels of redrawing a reader will allow — one camera at 17.6 px, two at 11.1, three at 4.6. The count is real: a picture built with two groups hidden in it gives back exactly those two, anywhere between nothing and 47.0 px. What it is not is a property of the picture alone, and the floor under it belongs to the hand — a one-camera drawing made by a hand that scatters its far corners by σ splits below about 0.57σ, where the 0.57 belongs to that hand and runs from 0.17 to 1.34 across a hundred and twenty of them.

choices · Divergence
axis — the ground line, fixed pointwisecentreaxisthe ground linefixed pointwisecentre(0.150, 4.020)height + distanceratio-1.481481−distance / heightheight × aspect = 2.592, and neither aloneeye recovered to 4.8e-12 mmeye 1.62 m up, 2.40 m backthree numbers back to the eye: 2.2e-16 m

The marks name the place, not the height

Run the camera-recovery round trip on an anamorph — hand it the floor marks and ask for the eye back — and it returns the spot on the floor to eleven decimal places with nothing assumed at all. It does not return the height. What the marks fix is the product of the eye's height with the design's aspect ratio, and no amount of looking at the floor separates the two.

viewing · Anamorphrecovery
front view · 32 filledtop view · 36 filledside view · 36 filledthe largest solid with these views — 192 cellsand a solid with the same views — 64 cells6 × 6 × 6 cellsthe three views are identical; the solids differ by 128 cells

Three views do not fix the solid

A stepped block on a six-cell grid draws a front, a top and a side view. So does a solid with a third of its material, and so does one with more than the block has — 192 cells against 64, every filled square in all three views identical. The drawing office's triple bounds a part between two solids and does not determine it, and the gap runs to a factor of n.

parallel · Multiview
hull 3.1098 · object 2.9193 · floor 3.07498 outlinesthe notch survives all of them

The dish no outline reaches

An outline is a pair of numbers per direction and nothing more, so a shape built from outlines has two errors that behave differently. The part outside the convex hull falls as one over the square of the view count. The part inside a concavity is the same area at four views and at a hundred and twenty-eight, because no pair of supporting lines ever reaches into a bite.

metrology · Visualhull
00.50011.50210203040how many marks the fit was givenerror in the recovered translation direction, in degrees (powers of ten)flatwith depththe same reading error in both5 counts

An ambiguity is not an uncertainty

Eight marks to forty cuts a solid scene's pose error from 19.8° to 0.7° and leaves a flat one at 48°. The two failures look identical from inside — a confident answer, a residual at the floor — and they respond to opposite remedies, so telling them apart is worth more than either measurement.

twoviews · Critical
020040060018202224along-roll scale measured beside the point (px of paper per metre of ground)separation of the point's own two drawings (px)5 m out10 m out20 m out40 m out80 m out120 m out6 points, one bendradius read back to 0e+0 m

A scroll can be asked its own radius

The two marks a bend leaves separate exactly. The along-roll scale alone fixes the angle in the disparity, so one point and a neighbour at its depth give back the radius and the depth in closed form — 200 m and 40 m returned to a part in 10⁹, with no search. The two answers are not equally held: a scale read one per cent too large under-reads the depth by one per cent and over-reads the radius by tan(φ − α)/α, which is 50 for a point ten metres from a five-hundred-metre bend. And a painter who evens the scale out by eye reports a gentler bend, never a bend that was never there.

scroll · Pushbroom
three eyes, flat ground1.6 m4 m11 mone eye, stepped groundground 9.4 mground 7.0 mground 0.0 mone eye, 11 m99 ground samplesagree to 6e-14 px

The stations are also a staircase

A level eye draws a ground point on a row that depends on the eye's height and the ground's only through their difference. So a landscape drawn from three stations at 1.6, 4 and 11 m over flat ground is, to 6 × 10⁻¹⁴ px across ninety-nine samples, the same picture as one eye at 11 m over ground stepped 9.4 m and 7.0 m above the far plain. The station heights a reader recovers are recovered on an assumption the marks neither supply nor test — and what would decide between the two readings is a cliff 103% of a band's own height, which is exactly where the convention lays its mist.

scroll · Threedistances
2468-1-0.50000.5001position along the family the two pictures leave freehow far the solid is from the scene's own shape (stretch ratio)the sceneevery member redraws both pictures to 7e-16 mthe ambiguity is a family of solids, not a tolerance

What two parallel views leave free

Two perspective pictures give shape and no size. Two parallel pictures do not even give shape: the metric upgrade has six unknowns, two views supply six equations, and the system comes out rank five — a one-parameter family of solids that redraw both pictures to 7e-16 m. A third view closes it, and the mirror image survives every count.

parallel · Parallelrecovery
the horizon, shared by every level stationthe shadows run tothe light, 22.0° belowband 1, station 1.6 mband 2, station 4 mband 3, station 11 mthree stations, one sun, 22° up and 30° round0e+0 px apart

A shadow decides which landscape it is

One sun over a several-station landscape is one sun: every band images the light's direction and its shadows' at the same page point, and each reports the altitude as 22.000000°. A shadow is still dislocated at the join, and the reading that could not be settled by any ground point is settled by one — because a ray crossing the seam has a riser to descend that the flat reading does not give it, and the two tips land 5.94 m and 11.7 px apart.

scroll · Threedistances
the true planefirst number of the plane, from the true valuesecond44 points in front: 31.9 % of the slicethe third number held at its true value

Seeing the scene fences in the plane at infinity

A reconstruction made without the calibration does not know which of its planes is infinitely far away. Requiring every point to lie in front of both cameras rules out every candidate that would tear the courtyard, and what is left is a convex region — 32 per cent of a generous slice — that always holds the true plane and never shrinks to it.

twoviews · Cheirality
0.2000.4000.6001.8022.20range, mheight above the plane of the eyes, mhorizontalverticalat 2 m, 40 cm off the plane79° between them

The second disparity cuts cells

A point off the plane of the eyes has a vertical disparity as well as a horizontal one, and quantising both, on an 86,400-point lattice of a room, gives 7,663 labels where one coordinate gives 179 — a count that belongs to the lattice rather than the room, as the essay after this one found. The gain is entirely vergence's — two eyes looking straight ahead have no vertical disparity at all, exactly — and it is largest where the first reading is already finest: 60.8 in the near metre and 3.7 in the far band.

depth · Disparity
the shared vanishing line, 22.0°four surfaces, four turns, one camera6e-12 px apart

A vanishing line with a slope in it

Turn the plane about the view direction and no family of any surface's edges is level; each vanishing line acquires a slope, and a group must agree about two numbers rather than one. The count does not change character — a hand of four pixels costs the test 3.00 px at no slope and 3.27 at thirty-eight degrees of it. What the slope does expose is the redraw: holding each far edge at its drawn height charges 0.95 px to a picture one camera really took.

choices · Divergence
plan: two different solidsthe picture, unmoved to 0e+0 pxcabinetkernel (0.32, 0.32, -0.89)

What one oblique drawing shows

A parallel projection is a linear map from three dimensions to two, so it has a direction it throws away. Slide any point along that direction and its mark does not move — which makes every parallel drawing the drawing of a three-parameter family of solids, and the depth scale a convention chooses is one direction through the family rather than a boundary of it.

parallel · Oblique
the second picture0.650.710.431.421.6122 raised marks, disc size: the coefficient on the epipolek ∝ h/Z to 4e-14

A parallax length is a height over a depth

After a known plane's map, every raised mark's displacement points at the other camera's image, and its length carries the mark's height above the plane over its depth — but not as the ratio of lengths it looks like. That ratio departs from the point's own number by up to 45 per cent. Read as a coefficient on the epipole, the same length gives height over depth from the first camera to four parts in a hundred trillion, the same from every second picture.

twoviews · Planar
correct from 20 cm, at 160 mm wideresidual 2e-15 m · 1/sin up to 1.22

A shadow across a second object

A straight edge held in front of a lamp defines one plane, and the shadow's boundary is wherever that plane meets something. That turns a picture of a shadow into a measurement: a camera ray and a known plane meet in one point, and the object the shadow is falling on comes back out.

light · Curvedreceiver
horizonflat on the groundleaning 23.6°two circles, 6.4 m across, in planes 23.6° apartcorrect from 26 cm, at 160 mm widetwo poses, 23.6° apart, one picture

Two circles, one picture

A photographed circle leaves its own pose ambiguous, and not a little: two congruent circles in planes 23.6° apart draw the same conic to 1.1e-16 on normalised coefficients, both of them in front of the camera. On top of that the distance is free, so the family is two discrete poses each with one continuous parameter — and a plane one degree from either draws a conic 2.5e-4 away, which is what makes the agreement a measurement.

foundations · Fiveconic
plan: the apex has movedfree along the kernelisometricrecovered to 1e-16

The drawing that gives the solid back

One parallel view of a general point determines nothing: two equations, three unknowns, and the kernel is free. What closes it is not a second view but the correspondence — knowing which drawn edge runs along which world axis — and with it the whole solid comes back out of one drawing, exactly.

parallel · Ruler
one platethe otherthe hingedrawn at 94.3° · reads as 100.0° or 86.6°an orthographic drawing of a foldtwo readings, 13.4° apart

A drawn fold has a phantom

A Necker cube has two readings and so does a drawn fold, and the fold's second reading is not the supplement of the first. The drawing fixes each plate edge's component in the picture and leaves its component along the ray free up to a sign; a reflection identifies two of the four sign pairs, so there are exactly two plates — and a hundred-degree fold reads as eighty-seven as well.

parallel · Foldreversal
frontierthe transfer joins points 1.414 R apartone ball, two outlinesmeeting in exactly two points

Two outlines are two curves

The whole method of multiview drawing is the transfer line — a feature at a position in the front view is at the same position along that axis in the top view. On a flat-faced solid the feature is a vertex and the rule is exact. On a ball the two views draw two different great circles, meeting in exactly two points, and the transfer line joins places that are √2 radii apart.

parallel · Parallelquadric
02.5057.5010020406080how far the floor is folded from flat (°)what a 10 cm error in the height leaves (mm)two degrees: 0.58 mma design 1.4 m wide, the eye 1.62 m up, the crease at 1.2 mflat: 1.5e-13 mm · 90°: 9.3 mm

A fold names the height

A pavement anamorph’s marks fix where the reader must stand and leave how tall they are entirely free — every height explains the marks exactly, to the last bit. Put one crease in the floor and the freedom is gone, because two degrees of fold makes a ten-centimetre error in the height leave six tenths of a millimetre, and a right angle makes it nine.

viewing · Anamorphrecovery

Named alongside it

The objects these essays reach for when they reach for this one.

DemonstrationBaselineConditioningDegeneracydegrees of freedomDrawing systemFree parameterinstrument limitResidualFundamental matrixParallel projectionTriangulation

All concepts