#48: Scikit-learn 45:Supervised Learning 23: Image augmentation in Python

preview_player
Показать описание
The video discusses the intuition for image data augmentation followed by coding in Python.

Timeline
(Python 3.8)

[Note: There is a lag in video and audio.]

00:00 - Outline of video
00:29 - What is image data augmentation?
01:39 - Why do data augmentation?
01:51 - Example of image augmentation
02:42 - Open Jupyter notebook
03:15 - Create a custom image augmentation function: read image
07:54 - Create a custom image augmentation function: rotate image
10:36 - Create a custom image augmentation function: shear
11:42 - * * * Note: Video lags audio * * *
13:33 - Create a custom image augmentation function: wrap
15:00 - Create a custom image augmentation function: noise
15:52 - Create a custom image augmentation function: blur
16:40 - Create a custom image augmentation function: flip
17:17 - Create a custom image augmentation function: List of augmented images
18:10 - Create a custom image augmentation function: Lables for augmented images
18:40 - Plot augmented images
21:24 - Ending notes

Download data:

Code for image augment function:
Рекомендации по теме
Комментарии
Автор

Hi Nilesh,
Image augmentation was very interesting!
But when you manipulate the same dataset does it not introduce multicollinearity into the model and lead to overfitting of data?

kshitijdesai
welcome to shbcf.ru