Back to Search
Start Over
Super-resolution recurrent convolutional neural networks for learning with multi-resolution whole slide images
- Source :
- Journal of Biomedical Optics
- Publication Year :
- 2019
- Publisher :
- SPIE-Intl Soc Optical Eng, 2019.
-
Abstract
- We study a problem scenario of super-resolution (SR) algorithms in the context of whole slide imaging (WSI), a popular imaging modality in digital pathology. Instead of just one pair of high- and low-resolution images, which is typically the setup in which SR algorithms are designed, we are given multiple intermediate resolutions of the same image as well. The question remains how to best utilize such data to make the transformation learning problem inherent to SR more tractable and address the unique challenges that arises in this biomedical application. We propose a recurrent convolutional neural network model, to generate SR images from such multi-resolution WSI datasets. Specifically, we show that having such intermediate resolutions is highly effective in making the learning problem easily trainable and address large resolution difference in the low and high-resolution images common in WSI, even without the availability of a large size training data. Experimental results show state-of-the-art performance on three WSI histopathology cancer datasets, across a number of metrics.
- Subjects :
- Paper
Computer science
ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION
Biomedical Engineering
Image processing
Context (language use)
01 natural sciences
Convolutional neural network
Imaging
Data modeling
Machine Learning
010309 optics
Biomaterials
Neoplasms
Image Interpretation, Computer-Assisted
convolutional neural networks
0103 physical sciences
Humans
image super-resolution
Image resolution
Microscopy
business.industry
Histological Techniques
Digital pathology
Pattern recognition
Image segmentation
Atomic and Molecular Physics, and Optics
whole slide imaging
Electronic, Optical and Magnetic Materials
Transformation (function)
pathology
Neural Networks, Computer
Artificial intelligence
business
Algorithms
Subjects
Details
- ISSN :
- 10833668
- Volume :
- 24
- Database :
- OpenAIRE
- Journal :
- Journal of Biomedical Optics
- Accession number :
- edsair.doi.dedup.....343fa24b741f2ac299b27677a49cb3d4