Download presentation
Presentation is loading. Please wait.
1
Advanced Image Processing
Spring. 2017 Dept. of Computer Science and Engineering
2
Image Processing Definition[Wikipedia]
Processing of images using mathematical operations by using any form of signal processing for which the input is an image, a series of images, or a video, such as a photograph or video frame; the output of image processing may be either an image or a set of characteristics or parameters related to the image.
4
Areas of IP Early processing Segmentation Descriptors
Filtering for noise reduction Enhancement Super resolution Feature extraction Segmentation Contour based Region based Video segmentation Descriptors Region Descriptors: color, texture Relational Description Image understanding Object detection Classification Image coding Vague boundary between IP and CV
5
Advanced Image Processing
Recommend prerequisites Digital Image processing Signal processing Probability theory Probability, Entropy, Bayes theorem Linear Algebra Matrix, Linear operators
6
Topics in Advanced Image Processing
Fourier Transform with its Applications Wavelet Transform with its Applications Image Pyramid and Multi-resolution Other Image Transforms Partial Differential Equations for Image Processing Image Compression Image Enhancement Techniques Contrast Enhancement Noise Reduction Super Resolution Local Image Descriptors Colors Image Restoration Mathematical Morphology Active Contours Snake and Level Sets
7
But recently Because of machine learning
Some topics are out-of-date or replaced by deep learning in part. Local descriptors Visual descriptors of MPEG-7 becomes meaningless. Mathematical morphology Partial differential equation techniques Image transforms Image enhancement … Now the changes is still going on.
8
Changes caused by … Machine Learning Deterministic models
Convolutional neural nets Discriminative models in deep learning Probabilistic models Bayes theory, Graphical models Generative models in deep learning
9
Strategy to Proceed IP class
Fundamentals should be covered at first by lecture. Some topics need to be briefly introduced and to be covered by recent papers. Milestone papers should be dealt with. But too many papers are poured out. You have to choose by yourself, read, and present in the class.
10
Not strictly organized class
Too many topics Selected topics that depends on the interests of the lecturer Rapidly changing area Hard to catch up completely Don’t know exactly what is next I am interested in advanced early processing. So I am focused on that topic in the class.
11
Environment of Practice
So many open sources floating around the web. OpenCV Github Libraries for Deep Learning Programming Languages Python Matlab C or Java
12
Grading Policy Midterm Exam 30% Assignment and computer projects 40%
Project Report Objective Procedure Results Discussion References Assignment will be given whenever a topic is finished. Presentation after 2/3 of the semester 20% Attendance in the class 10%
13
Note Lecture material for coming Thursday will be uploaded no later than Tuesday in the week on You should bring the downloaded copy in the class. Try to follow the lecture plan on the web Next week: Fourier transform and its applications
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.