BriefGPT.xyz
Feb, 2025
随机共振改善深度学习模型中低对比度图像的检测
Stochastic Resonance Improves the Detection of Low Contrast Images in Deep Learning Models
HTML
PDF
Siegfried Ludwig
TL;DR
本研究解决了传统深度学习模型在低对比度图像分类中表现不佳的问题。通过在LSTM递归神经网络中引入控制噪声，发现随机共振现象能够显著提高对弱信号的检测能力。结果表明，该方法有潜力改善图像分类性能，尤其在低对比度的情况下。
Abstract
Stochastic Resonance
describes the utility of
Noise
in improving the detectability of weak signals in certain types of systems. It has been observed widely in natural and engineered settings, but its utility in <
→