Instrumented People, Places and Things: the UIC Innovation Lab, 2012

Blue points

The depth frame: a point for each pixel of 640 × 480 with a reading, in the viewer's blue. Points in, below, gives a gray by depth instead; on the view, the keys 1 and 3 do the same, as they did in 2012. Its key 2, the stored color, has nothing to draw here: this frame was saved without color.

Colored lines

A tracker id: a dot for each reading of its center of mass, joined by thin lines. 1 red, 2 green, 3 blue, 4 yellow, 5 purple, any other orange.

Large dots

Where each id is at the replay's time. Added in 2026: the viewer drew a file whole. At the end of the replay the picture is the viewer's.

Yellow lines

The Kinect's field of view, from where it is placed.

The models

The viewer's, in its file's colors and lit as the sketch lit it. The office, the open house's, with its baked textures; drawing there is added. Both on the viewer's black ground, in light and dark.

Tracks

Under the view, when each id has a reading. Added in 2026. Click to set the time.

16 Jul

The depth frame at 12:36. The positions from 13:15, for 8 min 40 s. A note beside the sketches, dated 19 July, calls these “the fixed data gathered from the meeting”.

17, 20 Jul

Exports of the model, the last, at 14:27 on the 20th, the one the viewer loads.

24 Jul

The viewer as it survives, copied at 10:40.

3 Aug

The viewer's saved alignment, at 11:31. TJ's placement stands in for it here.

Left out

Other frames and recordings, from 22 July to 3 August, are not in this draft.

In the viewer's model. The viewer's model loader flips y, and the viewer scales the model by 12. The frame and the positions go in by TJ's placement. TJ, 4 November 2026: “The UIC point cloud is off, although it is close. The corner in the point cloud is the corner in the room containing the Kinect. So rotate 45 clockwise? Ish?” That room is the bay of two crossing partitions 8 ft high. The frame's clear corner goes on the corner where their faces meet, its walls along them and its floor on the model's floor, at true scale with nothing stretched. By the wall lines the turn is 50.7° clockwise: 51.8° by one wall, 49.6° by the other. It is about that corner, because turned about the Kinect the two corners would stay 62 in apart; so the Kinect moves 62 in, and stays in the bay. The whole world is then mirrored once for display, as Processing's screen showed it, so the Kinect's right is on the right.

The 2012 alignment. This placement is TJ's, and not what the viewer saved on 3 August 2012. That alignment stood the Kinect 30.7 in above the floor in the same bay, looking at a partition 12.6 ft away, and drew its data about 11% small.

In the office. The model goes in by the open house's rule: the same flip, its own turn and move. The frame goes in by a fit added in 2026: its floor on the floor, its clear corner on the room's north-west corner, its walls on the walls, at true scale with nothing stretched. The same mirror follows.

Walls

The north wall comes out at 212.5 in, the model's at 214. The west at 175.3 in, the model's at 174.

Past them

0 of 146,055 points above 40 in pass either wall by 6 in. The highest are at 107.5 in, under the 120 in ceiling.

Positions

4,730 of 4,731 in the room, the last within an inch; their middle 902 mm above the frame's floor, about where a center of mass would be.

The step

4.9 in into the room at 114.2 in from the corner, against the pier's 4.5 in at 108.5 in. The fit used neither.

What is off

The corner measures 87.2° where the model has 90°. The Kinect comes out at 72.4, 317.0, about 10 in past the partition and 7 in past the window wall, just behind the model's brick pier. Far depth that read long could do both: about 1% per meter of distance squares the corner, brings the step to 107.3 in, and moves the Kinect to 79.4, 310.3, inside that pier, about 13 in short of the room.

No file records the event at the Lab, or says when it was, and no photo or video of it has been found. UIC in a folder's name says what a sketch was made for, not where a recording was made.

TJ's placement shows the frame in the Lab's model, and the frame itself was made in the office. In that bay its walls lie within 2 in of the partitions' faces and rise 11.5 in past their tops. The alignment the viewer saved in 2012 cannot be tested: nothing here was recorded with the Kinect where it puts it.

A person here is one point. An id is the tracker's label for as long as it holds on to something. Nothing says who anyone was or what they were doing, so this stops at observation, short of a model of activity.