Today’s Activities:
- Finished the YOLO project, did multiclass classification on web cam feed or pre-recorded videos.
Homework:
Watch the following video and write your answers to a file called Jun4_CNN_HW1.txt. Upload to google drive when you are done:
- What does CNN stand for, and what type of data is it especially good at working with?
- Why is it hard for a regular neural network to understand images?
Hint: Think about how many pixels an image has. - What is a filter/kernel in a CNN?
Explain it in your own words. - What kind of “features” might early CNN filters detect in an image?
- What is a feature map?
Notes:
You can reach me at ddjapri@ayclogic.com.
All class notes can be found here.