FFT-based Selection and Optimization of Statistics for Robust Recognition of Severely Corrupted Images

arXiv:2403.14335v1 Announce Type: new
Abstract: Improving model robustness in case of corrupted images is among the key challenges to enable robust vision systems on smart devices, such as robotic agents. Particularly, robust test-time performance is imperative for most of the applications. This paper presents a novel approach to improve robustness of any classification model, especially on severely corrupted images. Our method (FROST) employs high-frequency features to detect input image corruption type, and select layer-wise feature normalization statistics. FROST provides the …

abstract agents applications arxiv case challenges cs.cv devices fft images improving key model robustness novel optimization paper performance recognition robotic robust robustness smart statistics systems test the key type vision