CS588: Deep Learning based Image Search (Spring 2024)

Instructor: Sung-eui Yoon

When: 10:30-12:00 on Mon. and Wed.
Where: Lecture room 3444, Information Science and Electronics Bldg (E3)
First class: Feb. 26 (Mon)
Textbook: In-class handouts and ongoing draft (web), ongoing draft (pdf) on image search
Board: KLMS
Question Page: Question Submission
Paper Submission Page: Paper Summary Submission (before every Mon. class)

Course overview

We extract feature points between two similar images and match them, followed by overlaying them together based on the matched points in the bottom row.

Thanks to rapid advances of digital camera and various image processing tools, we can easily create new pictures, images, and videos for various purposes. This in turn results in a huge amount of images in the internet and even in personal computers. For example, flickr, an image hosting website, contains more than five billion images and flickr members update more than three thousands image every minute.

These huge image databases pose numerous technical challenges in terms of image processing, searching, storing, etc. In this class we will discuss various scalable techniques for web-scale image/video databases and novel applications that can utilize such data.

In summary, what you will get at the end of the course:

Broad understanding on image/video retrieval techniques
In-depth knowledge on recent methods that can handle web-scale data
Study novel applications that utilize web-data

What you will do:

Choose and present a few papers from recent conferences.
Final project: come up with your own idea related to the topic, (optionally) implement it to improve the state-of-the-art techniques
Mid-term exam: reviewing basic image retrieval methods

Lecture schedule (subject to change)

Date	Topics and slides	Related material(s)
Feb. 26 (Mon)	Overview on the course and course policy	Lecture Video
Feb. 28 (Wed)	Classical Keypoint Localization
Mar. 4 (Mon)	Scale Invariant Region Selection and SIFT
Mar. 6 (Wed)	Deep Learning based Image Search	Programming Assignment 1
Mar. 11 (Mon)
Mar. 13 (Wed)	Re-Ranking and Inverted Index
Mar. 18 (Mon)	Hashing Techniques
Mar. 20 (Wed)	Person Re-identification
Mar. 25 (Mon)	Pixel Retrieval	Programming Assignment 2
Mar. 27 (Wed)	Diffusion for Objects Retrieval
Apr. 1 (Mon)
Apr. 3 (Wed)	Applications of Adversarial Attacks on Matching-based Algorithms
Apr. 8 (Mon)	Optical Flow
Apr. 10 (Wed)	No class due to the general election
Apr. 15 (Mon)	No class (midterm week)
Apr. 17 (Wed)	Midterm Exam
Apr. 22 (Mon)	Paper Presentation I: 1. Kyubeom Han 2. Sheikh Shafayat
Apr. 24 (Wed)	Paper Presentation I: 1. FILIPPO MOMENTE 2. Suhyeon Ha
Apr. 29 (Mon)	Paper Presentation I: 1. Jinhwan Seo 2. Jumin Lee
May. 1 (Wed)	Midterm Project Presentation: 1. T1 (Kyubeom Han, Jinhwan Seo) 2. T2 (Sheikh Shafayat)
May. 6 (Mon)	No class due to the substitute holiday
May. 8 (Wed)	Midterm Project Presentation: 1. T3 (FILIPPO MOMENTE) 2. T4 (Suhyeon Ha, Jumin Lee)
May. 13 (Mon)	No class due to ICRA attendance
May. 15 (Wed)	No class due to Buddha's birthday
May. 20 (Mon)	Paper Presentation II: 1. Sheikh Shafayat 2. FILIPPO MOMENTE
May. 22 (Wed)	Paper Presentation II: 1. Jinhwan Seo 2. Jumin Lee
May. 27 (Mon)	Paper Presentation II: 1. Kyubeom Han 2. Suhyeon Ha
May. 29 (Wed)	Reserved
Jun. 3 (Mon)	Final Project Presentation: 1. T1 (Kyubeom Han, Jinhwan Seo) 2. T2 (Sheikh Shafayat)
Jun. 5 (Wed)	Final Project Presentation: 1. T3 (FILIPPO MOMENTE) 2. T4 (Suhyeon Ha, Jumin Lee)
Jun. 10 (Mon) Jun. 12 (Wed)	Reserved (Final exam period)

Student presentations and reports

For your presentations, please use the this powerpoint template; paper presentation guideline is available.
For your final report, please use the this latex template

Additional reference materials and links

Computer vision resources (papers, videos, code, datasets, etc.):

CVPapers, Vision talk videos
Video lectures:
Multimedia Information Retrieval
VLFeat: contains popular computer vision algorithms including SIFT, MSER, k-means, hierarchical k-means, agglomerative information bottleneck, and quick shift

Paper search:

Acknowledgements: The course materials are based on those of Prof. Fei-Fei Li, Stanford. Thank you so much!

Copyright 2024. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the author.

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.