Improving Realistic Worst-Case Performance of NVCiM DNN Accelerators Through Training with Right-Censored Gaussian Noise.
Zheyu YanYifan QinWujie WenXiaobo Sharon HuYiyu ShiPublished in: ICCAD (2023)
Keyphrases
- gaussian noise
- worst case
- training process
- denoising
- noisy images
- signal to noise ratio
- noise level
- image restoration
- noise removal
- salt and pepper noise
- noise model
- jpeg compression
- noise suppression
- impulse noise
- white noise
- median filtering
- image analysis
- median filter
- training set
- lower bound
- noise variance
- training data
- image denoising
- high dimensional data
- higher order
- upper bound