Learning Continuous Image Representation with Local Implicit Image Function
This work provides a novel continuous image representation, which is significant for researchers working on image processing and generation, especially in super-resolution and handling varied image sizes.
This paper proposes Local Implicit Image Function (LIIF), a continuous image representation that takes image coordinates and local deep features to predict RGB values. Trained via self-supervised super-resolution, LIIF can represent images at arbitrary resolutions, even extrapolating to 30x higher resolutions than seen during training.
How to represent an image? While the visual world is presented in a continuous manner, machines store and see the images in a discrete way with 2D arrays of pixels. In this paper, we seek to learn a continuous representation for images. Inspired by the recent progress in 3D reconstruction with implicit neural representation, we propose Local Implicit Image Function (LIIF), which takes an image coordinate and the 2D deep features around the coordinate as inputs, predicts the RGB value at a given coordinate as an output. Since the coordinates are continuous, LIIF can be presented in arbitrary resolution. To generate the continuous representation for images, we train an encoder with LIIF representation via a self-supervised task with super-resolution. The learned continuous representation can be presented in arbitrary resolution even extrapolate to x30 higher resolution, where the training tasks are not provided. We further show that LIIF representation builds a bridge between discrete and continuous representation in 2D, it naturally supports the learning tasks with size-varied image ground-truths and significantly outperforms the method with resizing the ground-truths.