Researchers have published a paper on arXiv detailing the stability of scale-space metrics for comparing images. The study focuses on metrics derived from Gaussian scale-space representations, which have connections to Besov spaces and Wasserstein distances. The paper quantifies the metrics' resilience to geometric deformations and introduces rotationally-invariant versions for comparing tomographic projections. Additionally, it presents efficient algorithms for metric evaluation from finite samples and proves their robustness against additive noise, with findings illustrated by numerical experiments. AI
IMPACT This research contributes to foundational understanding in image analysis and metric stability, potentially impacting future AI models for computer vision tasks.
RANK_REASON The cluster contains a research paper published on arXiv.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
