| Responsible Department | Department of Food Science | ||||||||||||||||
| Research School | FOOD Denmark | ||||||||||||||||
| Course Dates | September 9th - 20th, 2013 | ||||||||||||||||
| Course Abstract | The course is designed to be an introduction to hyperspectral and multichannel images and their analysis in MATLAB environment. During the course, the students will learn how to extract the relevant information from their images as well as the fundamentals of the main multivariate analysis methods (chemometrics) applied. The course will be conducted by using in-house functions programmed in MATLAB. See the following link for further information (http://www.models.life.ku.dk/HYPERTools) | ||||||||||||||||
| Course Registration | To sign up for the course, please send an e-mail to Jeanette Venla Hansen (jvh@life.ku.dk) | ||||||||||||||||
| Deadline for Registration | September 1st, 2013 | ||||||||||||||||
| Credits | 7 (ECTS) | ||||||||||||||||
| Level of Course | PhD course | ||||||||||||||||
| Organisation of Teaching | - Contact Teaching: Slides. The students will follow all the exercises in their own computers. - Learning: Theoretical and practical exercises. - Educational approaches: Students with different educational background. | ||||||||||||||||
| Language of Instruction | English | ||||||||||||||||
| Restrictions | Maximum 15 participants. | ||||||||||||||||
| Course Content | |||||||||||||||||
| 1. Introduction 1.1. Types of images 1.2. General terminology 1.3. Outline of this course 2. Basic screening of images 2.1. Regions of interest 2.2. The multivariate approach 3. Pre-processing of images 3.1. Spatial pre-processing 3.1.1. Background removal 3.1.2. Spatial spikes. Dead pixels. Wavelets. Interpolation 3.2. Spectral pre-processing 3.2.1. Dead wavelengths. Wavelets. Interpolation 3.2.2. Spectral artefacts: De-noising, baseline removal, derivatives 3.3. Image compression 3.3.1. Wavelets 3.3.2. Multivariate compression 3.3.3. Spatial binning 4. Exploration of images 4.1. Principal Component Analysis. The MIA approach 4.2. Evolving factor analysis on images 5. Resolution of images. Multivariate Curve Resolution 5.1. MCR on images. 5.2. Constraints 5.3. Interpretation of results. 5.4. Augmented MCR 6. Regression models on images 6.1. Multivariate regression models 6.2. Validation of regression models on images 7. Segmentation 7.1. Definition of segmentation and differences with classification 7.2. Thresholding 7.3. Classification methods 7.3.1. Cluster analysis. Fuzzy clustering and K-means 7.3.2. PLS-DA 8. Topography 8.1. Features extraction from images. Area, diameter, excentricity, etc. 8.2. Domain statistics. Histograms 8.3. Fractals on images 8.4. The concept of homogeneity. Co-occurrence matrices | |||||||||||||||||
| Teaching and learning Methods | |||||||||||||||||
| - Contact Teaching: the basis of the course teaching will be done by presentations. Lectures in power-point plus examples/exercises with the computer. The students will be able to follow all the exercises in their own computers. - Learning: Apart from the exercises presented at class, the students will be given several exercises that the must solve for the report. - Educational approaches: This course is addressed to PhD student or advanced Master students. The educational background of the students will be different. | |||||||||||||||||
| Learning Outcome | |||||||||||||||||
| After the course the students will be able to apply basic multivariate data analysis to their own hyperspectral images using MATLAB as preferred platform. The course will focus on hyperspectral and multichannel images and their interpretation and analysis with multivariate methods (PCA, MCR, PLS, PLS_DA, etc). The methods treated will explicitly or implicitly cover the following application areas: classification, calibration, prediction, spectral resolution and interpretability of solutions. | |||||||||||||||||
| Course Literature | |||||||||||||||||
| Handouts and scientific papers provided during the course. | |||||||||||||||||
| Course Material | |||||||||||||||||
| Handouts and scientific papers provided during the course; scripts and source code provided during the course. | |||||||||||||||||
| Course Coordinator | |||||||||||||||||
| Jose Manuel Amigo Rubio, jmar@life.ku.dk, Department of Food Science/Quality and Technology, Phone: 353-32570 | |||||||||||||||||
| Course Fee | |||||||||||||||||
| - There is no course fee for any participant. - There is a fee of 1000 DKR for not attending when enrolled. | |||||||||||||||||
| Type of Evaluation | |||||||||||||||||
| Written individual report based on an examination assignment. The reports have to be handed in maximally one week after the last lecture and are evaluated and credited with PASS/FAIL by the course lectures. | |||||||||||||||||
| Work Load | |||||||||||||||||
| |||||||||||||||||
| Other Remarks | |||||||||||||||||
| Maximum 15 participants. Basic knowledge and experience with the MATLAB software. A PhD course in basic MATLAB programming and analysis is held one week before this course. Please contact José Amigo Rubio (jmar@life.ku.dk) for more information on this course. | |||||||||||||||||