醫學影像處理
Medical image processing
| 節 | 週五 |
|---|---|
2 09:00–09:50 | 醫學影像處理 EE117 3 節連堂 |
3 10:10–11:00 | |
4 11:10–12:00 |
* 根據陽明交大上課時間表所列
This course focuses on the fundamentals of medical image processing and its applications in machine learning. Students will systematically learn essential techniques such as image sampling and quantization, image enhancement, edge detection and sharpening, spatial domain filtering, morphological operations, image analysis, and segmentation. Special emphasis will be placed on feature extraction—a critical step for tasks like image segmentation and classification using machine learning. Through hands-on Python programming, students will bridge theory and practice to solve real-world problems in medical imaging. This course also serves as a foundational step toward deep learning in the following semester. *Course Objectives 1. Master fundamental image processing techniques: Understand and implement basic algorithms for sampling, quantization, enhancement, and filtering. 2. Develop strong feature extraction skills: Learn to extract discriminative statistical, morphological, and texture features from images for machine learning models. 3. Apply machine learning to medical images: Use classification and segmentation models to address real-world challenges in medical imaging and evaluate model performance. 4. Strengthen Python programming skills: Build complete image processing and machine learning pipelines using Python and key libraries (e.g., NumPy, OpenCV, scikit-image, scikit-learn). 5. Foster problem-solving mindset: Gain the ability to analyze new imaging problems, select appropriate techniques, and rapidly prototype solutions, laying a solid foundation for future work in deep learning.
Basic calculus, linear algebra, introductory programming, and Python proficiency.
課程中講解原理,部分內容將配合原理引導實作。請務必準備筆電至本課程。本課程需用到Python,若無程式基礎的同學,需自學基礎Python語法。 This course integrates both theoretical instruction and hands-on practice. Students are required to bring a laptop to each class. Python will be the primary language used. Students without prior programming experience are expected to self-learn basic Python syntax before the semester begins. *Learn coding with Python Suggested Reading: [1] https://www.halvorsen.blog/documents/programming/python/resources/Python%20Programming.pdf [2]https://assets.openstax.org/oscms-prodcms/media/documents/Introduction_to_Python_Programming_-_WEB.pdf *Lab Sessions and TA Support All labs require a final worksheet report and executable code. During lab sessions, the professor or TA may ask students to demonstrate their code and explain their analysis, and provide support during practical exercises. *Final Project Requirements: The project will integrate theoretical knowledge and practical skills taught during the semester and be implemented in Python. (1)Project tasks, datasets, and evaluation criteria will be published on the course website. (2)Each student/team must submit complete source code, a written report, and deliver an oral presentation.
1. Class Participation (worksheets and assignments): 40% 2. Midterm Exam: 15% 3. Final Project: 45% (Oral Presentation+report:15%, Code/Performance:30%)
| 週次 | 主題 |
|---|---|
| 第 1 週 | Course Introduction Understand course logistics, evaluation, and final project expectations. |
| 第 2 週 | Digital image fundamentals Overview of medical‐imaging workflow Digital image representation (pixels, resolution, dynamic range) |
| 第 3 週 | Image Enhancement(I) - Intensity Transformations - Histogram Processing & Contrast Enhancement |
| 第 4 週 | Image Enhancement(II) Segmentation Basics-thresholding (global, Otsu, adaptive) |
| 第 5 週 | image Filtering- Smoothing & Sharpening - Convolution - Smoothing filters & noise reduction trade‐offs - Derivative filters for edge enhancement |
| 第 6 週 | Suspension of classes(國慶日) |
| 第 7 週 | Edge Detection Fundamentals-Understand gradient‐based edge operators (Sobel, Prewitt...), canny detector - First & second derivative operators - Non‐maxima suppression & hysteresis - Edge linking & basic contour extraction |
| 第 8 週 | Suspension of classes(光復節補假) |
| 第 9 週 | Morphological Image Processing -dilation, erosion, opening, and closing - Binary morphology basics - Structuring elements design - Gray‐scale morphology - Connected component analysis ‐ Leverage morphological operations for noise cleaning and structure enhancement. |
| 第 10 週 | Midterm |
| 第 11 週 | Feature Extraction - statistical descriptors (mean, std, skew...) of ROIs ‐ shape features(area, perimeter, compactness, ellipticity...) - texture features (HoG, GLCM, LBP, Gabor...) |
| 第 12 週 | Classical Machine Learning (I) - a brief of ML ‐ classifiers |
| 第 13 週 | Classical Machine Learning (II) Performance Evaluation & Model Explainability |
| 第 14 週 | End‐to‐End Image Classification, Segmentation Pipeline |
| 第 15 週 | Project |
| 第 16 週 | Project Presentations ‐ Present and critique class projects. ‐ Identify limitations of classical methods and motivation for CNNs. |
Lecture Notes Reference: 1. Rafael C. Gonzales, Richard E. Woods, "Digital Image Processing", 4th Edition, Pearson Education, 2010. https://www.cl72.org/090imagePLib/books/Gonzales,Woods-Digital.Image.Processing.4th.Edition.pdf 2. Christopher M. Bishop, "Pattern Recognition and Machine Learning", 2006. https://www.microsoft.com/en-us/research/wp-content/uploads/2006/01/Bishop-Pattern-Recognition-and-Machine-Learning-2006.pdf
- 時間
- Appointment by email
- 聯絡方式
- chiayenlee@nycu.edu.tw