Predicting Rotation
Explore how to use rotation prediction as a pretext task in self-supervised learning. Understand how neural networks classify rotated images into four categories and how this helps models learn useful features without labeled data. Gain practical skills by implementing rotation transformation, generating pseudo labels, and training a ResNet18 model for rotation classification.
We'll cover the following...
We'll cover the following...
Rotation prediction
As illustrated in the figure below, the critical idea of rotation-based self-supervision is to rotate an input image and then ask the neural network to predict which rotation degree was applied to the picture. Generally, this pretext task is posed as a