Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,780 --> 00:00:04,500
In this video, we're going to be taking a look at another deepfake detector.
2
00:00:04,740 --> 00:00:06,330
And this one's called deep where.
3
00:00:07,110 --> 00:00:14,370
And if you're going to browse to it from the web, you can find it over deep where dot a I know this
4
00:00:14,370 --> 00:00:16,470
is another great tool again we could use as.
5
00:00:18,340 --> 00:00:25,270
Through the Web browser, which we're going to do in a moment, there's also a Google Play app that
6
00:00:25,270 --> 00:00:26,080
you can install.
7
00:00:27,340 --> 00:00:32,440
And also another thing I like is here's a GitHub repository for this, and we could see the deepfake
8
00:00:32,440 --> 00:00:40,360
scanner here and also some of the additional modules that you could do, like face face detection recognition
9
00:00:40,360 --> 00:00:41,080
library.
10
00:00:42,010 --> 00:00:47,200
And this is great if you want to, if you're building up your own machine and you want something kind
11
00:00:47,200 --> 00:00:49,150
of contained in your own little toolbox.
12
00:00:51,430 --> 00:00:56,680
But again, we're going to take a look at the Web browser version of this, so it's pretty easy, just
13
00:00:56,680 --> 00:01:00,100
like the other one, if we go to go to scanner here.
14
00:01:01,440 --> 00:01:07,890
We could either enter or you are ill here or we could do skin, and I'm just in, or you can click this
15
00:01:07,890 --> 00:01:11,460
little button here to upload a file, so I'm going to just upload a file.
16
00:01:11,470 --> 00:01:13,080
This is one of the videos I created.
17
00:01:16,220 --> 00:01:17,930
We give them a moment to upload here.
18
00:01:19,640 --> 00:01:25,340
And it's going to go through the skin process now, depending how large that video is, I ran this last
19
00:01:25,340 --> 00:01:28,700
night is it's going to depend on how long it takes.
20
00:01:28,880 --> 00:01:35,390
Generally, this is a if we can see it as about a twenty six point seven MEG file, it took a couple
21
00:01:35,390 --> 00:01:35,870
of minutes.
22
00:01:36,320 --> 00:01:37,760
I would say less than five minutes.
23
00:01:37,760 --> 00:01:39,110
Much less than five minutes.
24
00:01:39,290 --> 00:01:41,210
More like two minutes or so.
25
00:01:43,090 --> 00:01:44,500
To do the detection.
26
00:01:45,780 --> 00:01:52,800
And we could see the user time and date here, and we could see the likelihood that this is a deep fake
27
00:01:52,800 --> 00:01:56,550
and no fix found for this particular one because, well, there's.
28
00:01:57,210 --> 00:01:59,730
It was actually going through a tutorial for.
29
00:02:01,060 --> 00:02:01,960
Or something else?
30
00:02:02,350 --> 00:02:08,770
But again, it's pretty easy we could see if it found a face, we could see also on this little this
31
00:02:08,770 --> 00:02:10,790
little short here.
32
00:02:11,560 --> 00:02:19,480
How likely it is a a deep fake scan and we could see other things like duration, resolution, frame
33
00:02:19,480 --> 00:02:23,620
rate, whatever Kotecha was recorded in and also same for the audio.
34
00:02:24,610 --> 00:02:26,530
So again, pretty easy tool to use.
35
00:02:27,700 --> 00:02:35,110
It's currently free, and if you're looking for something to do a scan on deepfake, you could try to
36
00:02:35,110 --> 00:02:37,060
take a look at the video itself.
37
00:02:37,060 --> 00:02:39,370
See if there's anything unusual about it.
38
00:02:39,850 --> 00:02:40,570
Pause it.
39
00:02:41,600 --> 00:02:43,970
Broken down frame by frame, if you need to.
40
00:02:45,050 --> 00:02:52,190
But again, this is another easy tool to use to try to protect if that is in fact a deepfake video anything
41
00:02:52,200 --> 00:02:55,130
we could find this right deep where dot A.I..
42
00:02:55,280 --> 00:02:56,000
Thanks for watching.
43
00:02:56,090 --> 00:02:56,360
We'll see.
4103
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.