TEMImageNet and AtomSegNet Deep Learning Training Library and Models for High-Precision Atom Segmentation, Localization, Denoising, and Super-resolution Processing of Atom-Resolution Scanning TEM Images.
Ruoqian LinRui ZhangChunyang WangXiao-Qing YangHuolin L. XinPublished in: CoRR (2020)
Keyphrases
- super resolution
- high precision
- image resolution
- low resolution
- higher resolution
- low resolution images
- high resolution
- image pyramid
- highest resolution
- deep learning
- resolution enhancement
- high resolution images
- denoising
- super resolved
- scanning devices
- test images
- multiresolution
- super resolution algorithm
- segmentation algorithm
- motion estimation
- segmentation method
- edge detection
- input image
- image restoration
- high quality
- single image
- image patches
- combining information from multiple
- object recognition
- image regions
- depth map
- natural images
- image classification
- multiscale
- image segmentation
- model selection
- three dimensional
- multiple images
- graph cuts
- level set
- face images
- high accuracy
- higher order
- machine learning