- Document Number:
20220027741
- Appl. No:
17/332696
- Application Filed:
May 27, 2021
- نبذة مختصرة :
Disclosed are an unsupervised learning method and an apparatus therefor applicable to general inverse problems. An unsupervised learning method applicable to inverse problems includes receiving a training data set and training an unsupervised learning-based neural network generated based on an optimal transport theory and a penalized least square (PLS) approach using the training data set, wherein the receiving of the training data set includes receiving the training data set including unmatched data.
- Assignees:
KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY (Daejeon, KR)
- Claim:
1. An unsupervised learning method applicable to inverse problems, comprising: receiving a training data set; and training an unsupervised learning-based neural network generated based on an optimal transport theory and a penalized least square (PLS) approach using the training data set.
- Claim:
2. The unsupervised learning method of claim 1, wherein the receiving of the training data set includes receiving the training data set including unmatched data.
- Claim:
3. The unsupervised learning method of claim 1, wherein the neural network includes: a first neural network configured to convert a first image obtained, as an input, from an intermittent Fourier spatial coefficient into a second image corresponding to a complete Fourier spatial coefficient; a Fourier transform unit configured to output a third image corresponding to the first image by applying a Fourier transform and an inverse Fourier transform to the second image; and a second neural network configured to discriminate between the second image and an actual image for the second image.
- Claim:
4. The unsupervised learning method of claim 3, wherein the neural network is trained via unsupervised learning based on a cyclic loss between the first image and the third image and an adversarial loss between the second image and the actual image.
- Claim:
5. The unsupervised learning method of claim 1, wherein the neural network includes any one of a neural network based on a convolution framelet and a neural network including a pooling layer and an unpooling layer.
- Claim:
6. The unsupervised learning method of claim 1, wherein the neural network includes: a first neural network configured to output a first deblurring image corresponding to the first microscopy image when the first microscopy image is input; a conversion unit configured to convert the first deblurring image into a second microscopy image corresponding to the first deblurring image using a point spread function; and a second neural network configured to discriminate between the first deblurring image and an actual image corresponding to the first deblurring image.
- Claim:
7. The unsupervised learning method of claim 7, wherein the conversion unit includes a linear convolution layer corresponding to the point spread function.
- Claim:
8. An image processing method, comprising: receiving a first image; and reconstructing the first image as a second image corresponding to the first image using an unsupervised learning-based neural network generated based on an optimal transport theory and a penalized least square (PLS) approach.
- Claim:
9. An unsupervised learning apparatus applicable to inverse problems, comprising: a receiving unit configured to receive a training data set; and a training unit configured to train an unsupervised learning-based neural network generated based on an optimal transport theory and a penalized least square (PLS) approach using the training data set.
- Claim:
10. The unsupervised learning apparatus of claim 9, wherein the receiving unit receives the training data set including unmatched data.
- Claim:
11. The unsupervised learning apparatus of claim 9, wherein the neural network includes: a first neural network configured to convert a first image obtained, as an input, from an intermittent Fourier spatial coefficient into a second image corresponding to a complete Fourier spatial coefficient; a Fourier transform unit configured to output a third image corresponding to the first image by applying a Fourier transform and an inverse Fourier transform to the second image; and a second neural network configured to discriminate between the second image and an actual image for the second image.
- Claim:
12. The unsupervised learning apparatus of claim 11, wherein the neural network is trained via unsupervised learning based on a cyclic loss between the first image and the third image and an adversarial loss between the second image and the actual image.
- Claim:
13. The unsupervised learning apparatus of claim 9, wherein the neural network includes any one of a neural network based on a convolution framelet and a neural network including a pooling layer and an unpooling layer.
- Claim:
14. The unsupervised learning apparatus of claim 9, wherein the neural network includes a first neural network configured to output a first deblurring image corresponding to a first microscopy image when the first microscopy image is input; a conversion unit configured to convert the first deblurring image into a second microscopy image corresponding to the first deblurring image using a point spread function; and a second neural network configured to discriminate between the first deblurring image and an actual image corresponding to the first deblurring image.
- Claim:
15. The unsupervised learning apparatus of claim 14, wherein the conversion unit includes a linear convolution layer corresponding to the point spread function.
- Current International Class:
06; 06; 06
- الرقم المعرف:
edspap.20220027741
Processing Request
No Comments.