Speckle-to-Speckle: Ultrasound speckle reduction technique using unsupervised deep learning
- Title
- Speckle-to-Speckle: Ultrasound speckle reduction technique using unsupervised deep learning
- Author(s)
- Dongkyu Jung
- DGIST Authors
- Dongkyu Jung ; Jaesok Yu ; Sangyoon Han
- Advisor
- 유재석
- Co-Advisor(s)
- Sangyoon Han
- Issued Date
- 2022
- Awarded Date
- 2022/08
- Type
- Thesis
- Description
- Ultrasound, Speckle, Unsupervised learning, Deep learning
- Abstract
- Table Of Contents
-
1. INTRODUCTION 1
2. METHODS 5
2.1. Speckle-to-Speckle: Unsupervised learning-based framework 5
2.1.1. Background 5
2.1.2. Speckle-to-Speckle network design 7
2.1.3. Training and inference 8
2.2. Experimental setup 9
2.2.1. Data acquisition 9
2.2.2. Implementation details 11
2.2.3. Evaluation metrics 11
3. RESULTS 15
3.1. Evaluation using training dataset 15
3.2. Evaluation using validation dataset 17
3.3. Evaluation using unspecified dataset 18
3.4. Calculation complexity 20
4. DISCUSSIONS 22
5. CONCLUSION 24
6. REFERENCE 25
요약문 27
- URI
-
http://dgist.dcollection.net/common/orgView/200000629151
http://hdl.handle.net/20.500.11750/16788
- Degree
- Master
- Publisher
- DGIST
- Related Researcher
-
-
Yu, Jaesok
- Research Interests Biomedical ultrasound/photoacoustic imaging and therapeutics; Medical image processing; Machine learning-based diagnosis and signal processing; Nano-theranostic platform
-
- Files in This Item:
-
There are no files associated with this item.
- Appears in Collections:
- Department of Robotics and Mechatronics Engineering Theses Master
Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.