Image Interpolation

Image interpolation aims to generate intermediate images between two given images, or to increase the resolution of an existing image, maintaining semantic consistency and visual quality. Recent research focuses on leveraging deep learning models, particularly diffusion models, to achieve this, often incorporating techniques like attention mechanisms, flow estimation, and novel resampling functions to improve accuracy and efficiency. These advancements are impacting various fields, from medical image analysis (improving segmentation and classification) to computer graphics (enabling realistic image morphing and video frame interpolation), by enhancing image quality and enabling new applications.