Other
FreeComputer Vision
No ratings yetEnglish
OtherMy Learning
Mark your own learning status. Opening a course link does not establish enrollment.
Loading saved status…
Korshub's take
From pixels to features
The course walks through how convolutional networks turn raw images into features. It starts with a first classifier built in Keras, then looks at the mechanics underneath: convolution, ReLU activations and max pooling. Stride and padding get their own treatment because they quietly decide how a network sees an image.
Beyond the basics
Later lessons move from designing a custom convnet to data augmentation and transfer learning, the techniques most practitioners reach for when training data is limited. Two lessons are built around Kaggle competitions, so you work with the same kind of real datasets that competition entrants face.
Who it is for
- Learners who want a hands-on path into convolutional networks.
- Anyone who wants to see a convnet built one layer at a time.
