Visualize and Paint GAN Activations
This work addresses the need for better interpretability and control in GANs for researchers and practitioners, though it appears incremental as it builds on existing GAN frameworks.
The paper tackled the problem of understanding and controlling GAN-generated images by correlating hidden layer activations with structures, enabling image generation from semantic segmentation maps without requiring such data during training.
We investigate how generated structures of GANs correlate with their activations in hidden layers, with the purpose of better understanding the inner workings of those models and being able to paint structures with unconditionally trained GANs. This gives us more control over the generated images, allowing to generate them from a semantic segmentation map while not requiring such a segmentation in the training data. To this end we introduce the concept of tileable features, allowing us to identify activations that work well for painting.