PAC-FNO: Parallel-Structured All-Component Fourier Neural Operators for Recognizing Low-Quality Images

arXiv:2402.12721v1 Announce Type: new
Abstract: A standard practice in developing image recognition models is to train a model on a specific image resolution and then deploy it. However, in real-world inference, models often encounter images different from the training sets in resolution and/or subject to natural variations such as weather changes, noise types and compression artifacts. While traditional solutions involve training multiple models for different resolutions or input variations, these methods are computationally expensive and thus do not scale in …

abstract arxiv cs.ai cs.cv deploy fourier image image recognition images inference low natural operators practice quality recognition standard train training type weather world