Building and training SRGANs to enhance the quality of images
Kailash Ahirwar (~kailash) |
Super-resolution Generative Adversarial Networks is a type of GAN which can enhance the resolution/quality of images. Enhancing the quality of images has many use-cases like:
- To recover old low-resolution images
- To automatically enhance the quality of the camera feed in video surveillance, images transferred over the Internet and television broadcasting
and many more.![enter image description here
In this talk, I cover the following topics:
- Introduction to SRGANs and its architecture
- Data collection and preparation
- Model creation in Keras and Tensorflow
- Model training and hyperparameter tuning
- Using the trained model to enhance the quality of images.
This talk will be a hands-on session and will provide a deep down introduction to SRGANs and training SRGANs. After the talk, attendees will be able to train their own SRGAN network from scratch. This talk is for deep learning researchers who are good with Generative Adversarial Networks and have trained GANs before.
- Proficiency in the Python programming language
- Basics of neural networks
- Experience of training GANs in Keras and Tensorflow
- Experience of working with Jupyter notebook
Kailash Ahirwar is a machine learning and deep learning enthusiast. He is the author of the book titled "Generative Adversarial Networks Projects", published by Packt publications. He has worked in many areas of Artificial Intelligence (AI), ranging from natural language processing and computer vision to generative modeling using GANs. He is a co-founder and CTO of Mate Labs. He uses GANs to build different models, such as turning paintings into photos and controlling deep image synthesis with texture patches. He is super optimistic about AGI and believes that AI is going to be the workhorse of human evolution.