A total variation scale-space and image resizing
Yedidya Hyams, Moshe Porat · 2016
In modern days, images are used over a large span of displays and captured using both professional and common cameras. All images are automatically pre-processed and are later on often re-sized. Re-sizing and some of this pre-processing reduces the image quality. We suggest a method for re-sizing images based on a new scale-space aproach which dynamically chooses a least destructive filter for carrying out the re-sizing. We show that our algorithm has less imperfections than other methods and suggest an expansion of the algorithm for pre-processing images. Our conclusion is that this method could obtain better images out of cameras without changing any hardware.