1
00:00:06,700 --> 00:00:12,859
welcome back in this video we're gonna learn about image based learning models and create

2
00:00:12,859 --> 00:00:19,579
a project that will sort two types of shapes into one another that can differentiate between two

3
00:00:19,579 --> 00:00:24,859
types of shape two types of shapes okay so first of all you have to log in and once you log in you

4
00:00:24,859 --> 00:00:34,289
will see this is this is screen so then you can go to projects and create a new project

5
00:00:34,289 --> 00:00:51,329
will name this project as shape sorter and the recognizing model will be images then we can

6
00:00:51,329 --> 00:01:00,439
create as you can see our shape sorter is created we can click on this model and click on train

7
00:01:00,439 --> 00:01:07,420
once you are on the training screen this is what you will see and in order to train you have to

8
00:01:07,420 --> 00:01:13,299
give the model labels so we're gonna add new label and we're gonna say that this

9
00:01:13,299 --> 00:01:27,299
label is triangle that this label is triangle and we're gonna add me another

10
00:01:27,299 --> 00:01:35,959
label and say that this label is square okay so once you have these two labels

11
00:01:35,959 --> 00:01:42,040
you can either add images from the internet you can add images from your

12
00:01:42,040 --> 00:01:47,560
webcam or you can add it or you can draw images we're gonna draw images in this

13
00:01:47,560 --> 00:01:51,900
example since this is an image of a triangle we can draw it as a triangle

14
00:01:51,900 --> 00:02:01,040
just make sure that these images are somewhat similar to a triangle and not

15
00:02:01,040 --> 00:02:06,719
entirely different from triangles if you make a mistake you can just reset and

16
00:02:06,719 --> 00:02:15,280
and try and draw try to draw it again okay so we will add 10 examples like these in each of

17
00:02:15,280 --> 00:02:24,240
these labels and try to keep our examples as consistent as possible okay okay so once you

18
00:02:24,240 --> 00:02:30,479
have examples make sure that you have 10 examples in each at least otherwise the training model

19
00:02:30,479 --> 00:02:37,599
won't work so once we have these 10 examples we can go back to project and learn and test the

20
00:02:37,599 --> 00:02:44,560
model and since i have already trained this model you can just click on train new machine learning

21
00:02:44,560 --> 00:02:50,080
model and it will automatically train this model but since i have all this model already available

22
00:02:50,080 --> 00:02:59,120
we can just test this model by two types we can test it via images by our webcam and we can test

23
00:02:59,120 --> 00:03:05,840
by drawing so we'll test by drawing since i don't have a webcam available so we can draw any shape

24
00:03:05,840 --> 00:03:14,639
and test see it's recognized as a square with 86 confidence we can try a triangle

25
00:03:18,080 --> 00:03:24,479
and let's see what happens so our as you can see that our model is working