-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathscenes.qmd
89 lines (54 loc) · 2.21 KB
/
scenes.qmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
# Scene Understanding {#sec-scene_understanding}
## Introduction
Scene understanding refers to the analysis of the whole scene: the
agents and objects in it, the relationships between them, the actions,
the place, \...
## A Few Notes About Scene Understanding in Humans
### Rapid Scene Recognition
Molly Poter and her experiments on rapid scene recognition.
"Later she discovered that complex visual scenes can be perceived and
understood much faster than anyone had previously recognized. She showed
that subjects can identify the gist of a scene from an astonishingly
brief presentation. Here Potter made innovative use of rapid serial
visual presentation (RSVP)."
FIGURE: show one experiment
### Neural Processing
The time it takes for us to recognize a scene is remarkable once we
relate it to the structure of the neural pathways and the time it takes
for a neuron to process its input. Simon Thorpe \...
Face area
Place fusiform area.
IT
### Remembering Images
Another important aspect of human perception is memory.
A computer vision system can compare two images directly, even at the
pixel level. A human can not do this because it can not look at two
images simultaneously in parallel. A human needs to look at each image
sequentially, and compare the two images from memory. Therefore, memory
is a crucial aspect of human visual perception.
change blindness
The world as an outside memory.
### Recognition in Context
Understanding a scene is an integrated process.
Some context illusions.
Out of context image.
Objects that belong to the same class have a common set of affordances,
but the affordance of an object might change depending on context. As an
example, if we use a card-box as a table, we might still call it *a box*
but we understand the affordance of the box as being very different to
its typical use.
Put the shoes in a place that is easy to find later.
## Scene Representations
Schemas
Frames
Movies, stories.
Non-linguistic representations?
Scene graphs
Symbolic-representations
## Taskonomy
Place classification
Recognizing locations
3D reasoning and decomposing the scene into separated objects.
Understanding the rules of support, parts and attachments between
objects.
## Concluding Remarks