基于未经训练的注意力U-Net的低采样和噪声强大的单像素成像
Optics express
|November 22, 2024
概括
这项研究引入了一个未经训练的注意力U-Net,以减少单像素成像 (SPI) 中的噪音. 该方法在较低的采样率下提高图像质量,扩大SPI.
科学领域:
- 光学是什么?光学是什么?光学是什么?
- 图像处理 图像处理
- 机器学习 机器学习
背景情况:
- 单像素成像 (SPI) 使用结构光和单像素探测器 (SPD).
- SPI在检测过程中受到白噪声的限制,降低了图像质量.
- 现有方法在降低噪音和采样率低方面存在困难.
研究的目的:
- 为了减少SPI中的噪音,使用未经训练的注意力U-Net.
- 为了在较低的采样率下实现高质量的成像.
- 提高SPI在噪音环境中的适用性.
主要方法:
- 将未经训练的注意力U-Net与SPI模型相结合.
- 利用注意力机制来突出显示图像特征并抑制噪音.
- 使用数值模拟和实验验证.
主要成果:
- 有效地减少不同级别的高斯白噪声.
- 与现有方法相比,图像质量优越,采样率低于10%.
- 在不需要预先培训的情况下展示了概括性.
结论:
- 拟议的方法在噪音条件下显著提高了SPI性能.
- 没有训练的注意力U-Net为低采样率的SPI提供了一个强大的解决方案.
- 这项工作扩大了SPI在复杂噪音环境中的潜在应用.
相关概念视频
Downsampling
872
When considering a sampled sequence with zero values between sampling instants, one can replace it by taking every N-th value of the sequence. At these integer multiples of N, the original and sampled sequences coincide. This process, known as decimation, involves extracting every N-th sample from a sequence, thereby creating a more efficient sequence.
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
872
Upsampling
745
Managing signal sampling rates is essential in digital signal processing to maintain signal integrity. A decimated signal, characterized by a reduced frequency range due to its lower sampling rate, can be upsampled by inserting zeros between each sample. This upsampling process expands the original spectrum and introduces repeated spectral replicas at intervals dictated by the new Nyquist frequency. To refine this zero-inserted sequence, it is passed through a lowpass filter with a cutoff...
745


