Page 269 - Ai_417_V3.0_C9_Flipbook
P. 269
Step 2: Data Augmentation
Data augmentation is the process of increasing the
amount and diversity of data. We do not collect
new data, rather we transform the already present
data. Data augmentation means increasing the
amount of data by adding copies of existing data
with small changes. The image given here does not
change, but we get data on the image by changing
different parameters like colour, rotation, flipping and
brightness. New data is added by slightly changing
the existing data.
In the above example:
● We apply flipping and rotation transformation to create variations of the original images.
● We also simulate occlusions such as objects partially blocking the view to train the model to handle obstructed
scenarios.
Step 3: Data Generation
Data generation refers to generating or recording data using sensors. Recording temperature readings of a building
is an example of data generation. Recorded data is stored in a computer in a suitable form.
In the above example, of self-driving car. Data acquisition is done for creating fake driving scenarios with different
road conditions, traffic patterns, weather, and lighting to cover many possible situations.
Task #Experiential Learning
Visualise that you are in a big, mysterious forest and searching for hidden treasure.
Write four observations you will be making for finding the treasure and categorise them under the
heads viz: data discovery, data augmentation and data generation.
Observations Categories
1. …………………………..……..………………………..……....……....…….. …………………………..……..………………………..……....……....……..
2. …………………………..……..………………………..……....……....…….. …………………………..……..………………………..……....……....……..
3. …………………………..……..………………………..……....……....…….. …………………………..……..………………………..……....……....……..
4. …………………………..……..………………………..……....……....…….. …………………………..……..………………………..……....……....……..
Data Literacy 267

