Original subtitles

1

Hello and welcome back to the course on computer vision in today's eternal we're going to talk about

2

another powerful feature of the essays the algorithm and that is how it predicts object positions.

3

So here we got another lovely photo which we took in Italy and these are just part of what's so small

4

boats which are parked for the day or for the night and they're going to sometime soon go maybe fishing

5

or whatever they're meant to do.

6

And as previously we're going to overlay a grid over this image and then for every single center on

7

this grid we're going to introduce a couple of rectangles and we're going to stick with our example

8

three rectangles like that.

9

And again as we discussed previously we're going to take each one of these rectangles and think of it

10

as a separate image and then within each rectangle the conventional neural network will ask the question

11

is there a bullet or is there not a bone basically lost about objects if you if you watch this video

12

this is from the promotional video for the course and you posit at some point you'll see that this boat

13

actually at one in one frame was recognized as an airplane or there was like a boat and an airplane

14

recognized in one here.

15

So it kind of shows that that's Lewis looking for all classes of objects at the same time.

16

So yeah it's a good verification that sometimes those errors happen it's just shows that maybe it can

17

confuse an object and this does look like a bit not like a nonstandard board because it's like a rubber

18

boat.

19

Yeah.

20

So basically each one of these squares each one of these rectangles they go say this rectangle can there's

21

so many of them you can barely see.

22

Let's say this right over here is going to ask is like just looking at it by itself forgetting about

23

everything else is going to ask is there any object inside here.

24

Can I detect any features of an object.

25

Yes or no.

26

And so in our case in that well after the training in the ideal scenario of these falling rectangles

27

will pick up features of bolts and so let's get rid of everything else.

28

So there you go you can see that these rectangles do overlap with our ground truth and therefore it's

29

fair to say that they will pick up the sort of features.

30

But again just a couple of things I wanted to point out here that during training this might not be

31

the case maybe this rectangle pick up the boat on this rectangle and this rectangle won't pick up a

32

boat.

33

But then through the years of training process they will know they will learn what the features of boats

34

are and how to pick them up and so on.

35

So that's the training that we got to lend.

36

Lengthy hard part in the detection of the network after it's trained up this is what will happen.

37

So we kind of like combining the concepts of training and application so it's important just to keep

38

in mind what exactly we're talking about.

39

So this is the ideal scenario what would happen during train and during application after the network

40

has trained.

41

And so let's have a look at one specific set of these three boxes for instance this one over here.

42

So the interesting thing is that these boxes have picked up features of a boat ride so they can see

43

maybe the front of the boat and this anchor line maybe some you know you can see that it's in the water.

44

They can see like this.

45

This part isn't just standing out meaning that's where the driver stands.

46

And so some of you can see more some of them can see less but nevertheless not all of them can see the

47

full image of the boat.

48

None of you can see the full ground truth of the boat.

49

And so the question is how can a box predict how can like one of these temporary or inferred boxes how

50

can it predict where exactly the full boat will be.

51

That's the question.

52

So if you like if you use your hand to just close that this part of the image.

53

How can you complete this boat without knowing where it actually is supposed to be complete without

54

knowing the full ground truth.

55

How do you know where it will be completed.

56

And that is another special part in the s is the algorithm.

57

It's actually not that complex when you think about it.

58

So the underlying principle is that we have the ground truth when making these predictions and training

59

the algorithms so now we're back to talking about training.

60

So when training the answer the algorithm we have these ground troops we have these actual boxes.

61

And so we can then let these boxes let the algorithm make its predictions of the algorithm based on

62

what it can see based on these individual boxes it can then make a prediction.

63

OK.

64

I think the boat looks like maybe maybe I'll say it looks like this.

65

So the full boat the full rectangle should be around like this size although it wouldn't do that because

66

it knows that there's no features a boat overboard in here.

67

Right so.

68

But for instance it could say like it knows that this rectangle has features of both.

69

Right.

70

So it doesn't know maybe the boat actually goes up up to up to the top here.

71

So it does use the overlapping.

72

So it is it makes it easier.

73

So like if you go if you go back it will use overlapping and where they are overlapping that's of course

74

is a high probability that chance that there is a boat where the two rectangles which are overlapping

75

is a hybrid.

76

It is a boat.

77

But nevertheless you can still make errors it can still maim like come to the assumption that the boat

78

is higher longer or wider or something like that but the way this is mitigated is again through training.

79

So at the end once it's made its predictions there will be a CERN.

80

CERN or box that it suggests for this body will say that the box star said this pixel the bottom left

81

corner is here on the bottom right corner is instead of saying here it might say we're here.

82

And then it will actually that's resolved that box that it creates will be compared to the ground truth

83

and then again will be an era which can be calculated and if you think of it in terms of coordinates

84

it can be calculated CHONE distance.

85

There's lots of mathematics going on in the background so we're not going to go into that right now

86

how exactly the error is calculated.

87

But you can envisage how the arrow would could be expressed in mathematical terms and that era would

88

be then propagated back up through the network in order to help update the weights.

89

And again we talked about all of this like the back propagation process and gradient descent in the

90

annex for a and then artificial neural networks and convolutional neural networks but then the conceptual

91

that's what is happening.

92

So the bottom line takeaway from this tutorial is that through training two things happen the first

93

thing is that what we discussed previously is that each box will learn to better classify objects or

94

better identify if it has objects inside there.

95

And in each box will use the ground truth to assess that.

96

So for instance this box will say that yes there is about where is this box or this box.

97

This this box we want it to say yes there is about.

98

And this box we don't want it to say that there is a bot and that's what'll happen to training and they'll

99

get better at that.

100

But the second thing that will happen is they will get better at identifying the exact final output

101

rectangle of that they should create.

102

So we want we want these this algorithm this is D to produce this exact rectangle which matches the

103

ground truth rather than a large direct eye or rather than a shorter rectangle.

104

So for instance here you can see there is always going to be some sort of error.

105

So here you can see that this rectangle because this isn't the ground truth right.

106

So at the beginning we discussed there isn't this isn't this isn't ground truth.

107

We assume this to be the ground truth but as you can see.

108

And then at that time I said that I wouldn't draw a better rectangle around any of them except I'm into

109

this.

110

But now if you look carefully you can see that this rectangle could be a bit narrower.

111

Right you could see that it could be a bit narrower here and this rectangle could also be a bit narrower.

112

So it's taking a little bit more so there's no I can there's no bones in these pixels over here on the

113

right.

114

There's no boat in these pixels on the left and maybe even this one could be a tiny little bit narrower.

115

So there we go.

116

That's how one second that's how the training happens.

117

And so that's another element of the training that is important to keep in mind so maybe if this algorithm

118

was trained along with the one that was actually applied to drill what we took as the ground maybe if

119

it was trained for a longer maybe this bullet would have been more precisely identified during the application

120

process.

121

But having said that this was a quite a tough challenge because as you can imagine the blow boats are

122

kind of like moving around they're floating with waves the waves and also that the drone is flying.

123

So this is actually a pretty good result for those circumstances anyway so that's the second part that

124

happens in the training.

125

This is slow as you can see slowly we're building up the different components of this deal.

126

So training is how to learn how to classify objects through the ground truth but also is the ground

127

truth to identify the correct widths and heights of those boxes.

128

And in the next tutorial we'll continue and talk about the next part component of that as is the.

129

And of course see you there and until next time in your computer vision.