Deep Multi-Scale U-Net Architecture and Noise-Robust Training Strategies for Histopathological Image Segmentation.
Nikhil Cherian KurianAmit LohanGregory VergheseNimish DharamshiSwati MeenaMengyuan LiFangfang LiuCheryl GilletSwapnil RaneAnita GrigoriadisAmit SethiPublished in: CoRR (2022)
Keyphrases
- multiscale
- image segmentation
- image noise
- scale space
- salt pepper
- training set
- greater robustness
- management system
- registration errors
- noisy environments
- random noise
- multiple scales
- deep architectures
- estimation error
- image structure
- real time
- deformable models
- image analysis
- noise reduction
- training and testing data
- texture segmentation
- gaussian noise
- feedforward artificial neural networks
- training process
- noise level
- region growing
- noisy data
- spectral clustering
- software architecture
- missing data
- segmentation algorithm
- active contours
- wavelet transform
- multiresolution
- image processing
- computer vision