Researchers have developed a new framework called pHVI-ISPNet to improve night photography rendering by addressing perceptual distortions and color bias. This RAW-to-RGB model utilizes specific refinements like RAW-domain feature processing and wavelet-based propagation to preserve detail. It also incorporates dynamic loss coefficients and a feature distribution loss term to ensure stable learning and color constancy across varying exposure levels. Evaluations show the approach achieves competitive fidelity and sets new state-of-the-art results in color difference and perceptual similarity metrics. AI
影响 Introduces a novel framework for enhancing night photography rendering, potentially improving image quality in low-light conditions.
排序理由 This is a research paper detailing a new technical approach to image rendering.
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
