Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,300 --> 00:00:00,990
In this video.
2
00:00:00,990 --> 00:00:04,530
We'll explain some of the tasks that are often performed by computer vision.
3
00:00:05,560 --> 00:00:07,480
First image classification.
4
00:00:08,310 --> 00:00:13,110
Image classification is the process of determining the category of an image based on the objects in
5
00:00:13,110 --> 00:00:13,230
it.
6
00:00:14,210 --> 00:00:17,210
The input image classification is an image with one object.
7
00:00:17,210 --> 00:00:21,380
What the output is, what object is in the input image with its probability value.
8
00:00:22,210 --> 00:00:23,950
There is also a localization.
9
00:00:24,750 --> 00:00:29,580
Localization perform classification coupled with determining the location of the object in the form
10
00:00:29,580 --> 00:00:30,840
of a bonding box.
11
00:00:31,830 --> 00:00:33,360
Next object detection.
12
00:00:34,190 --> 00:00:39,470
The main purpose of object detection is to predict the location of objects with bounding boxes and classify
13
00:00:39,470 --> 00:00:41,300
objects in each bounding box.
14
00:00:41,960 --> 00:00:45,770
The input to object detection is an image containing one or more objects.
15
00:00:46,430 --> 00:00:51,020
The output is the prediction of the location of the object with the bounding box and the classification
16
00:00:51,020 --> 00:00:53,120
of the object for each bounding box.
17
00:00:53,810 --> 00:00:55,400
Next Image segmentation.
18
00:00:56,190 --> 00:00:59,580
Image segmentation is a further extension of object detection.
19
00:01:00,210 --> 00:01:03,270
Object detection only predicts the bounding box of the object.
20
00:01:03,270 --> 00:01:06,540
While in image segmentation, we can find out the shape of the object.
21
00:01:07,430 --> 00:01:11,900
There are two types of segmentation instant segmentation and semantic segmentation.
22
00:01:12,710 --> 00:01:18,080
Instant segmentation recognizes an object's boundaries and labels its pixels with a different color.
23
00:01:19,290 --> 00:01:24,300
Semantic segmentation assigns a different color to each pixel in the image, including the background
24
00:01:24,300 --> 00:01:26,880
based on the results of the object classification.
25
00:01:28,150 --> 00:01:28,570
From YOLO.
26
00:01:28,570 --> 00:01:32,020
If you want to YOLO, physics can only be used on object detection.
27
00:01:32,780 --> 00:01:36,740
YOLO v seven can be used for object detection and instant segmentation.
28
00:01:37,510 --> 00:01:42,790
The YOLO v seven is also planning to make YOLO seven semantic segmentation in the near future.
29
00:01:43,120 --> 00:01:43,780
See you then.
2800
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.