Scribd
Upload
6 views10 pages

CMU Convex Optimization Course Overview

The document provides an overview of a Convex Optimization course taught by HyungGi Jo at Jeonbuk National University, including lecture details, grading criteria, and course materials. The course aims to teach problem formulation and solution using optimization tools, with a focus on practical applications such as correspondence matching in images. Key resources include a textbook by Boyd and Vandenberghe and lecture notes from CMU and Stanford.

Uploaded by

Anh Lê
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views10 pages

CMU Convex Optimization Course Overview

The document provides an overview of a Convex Optimization course taught by HyungGi Jo at Jeonbuk National University, including lecture details, grading criteria, and course materials. The course aims to teach problem formulation and solution using optimization tools, with a focus on practical applications such as correspondence matching in images. Key resources include a textbook by Boyd and Vandenberghe and lecture notes from CMU and Stanford.

Uploaded by

Anh Lê
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Convex Optimization

Course Overview

HyungGi Jo
Div. of Electronics
Jeonbuk National University

AI Robotics Lab.
Introduction
Convex Optimization
Feb. 09:00 ~ 12:00
공과대학7호관 523
강의자료
▪ Lecture Notes (CMU Note, Stanford EE364a)
▪ 모두를 위한 컨벡스 최적화
• [Link]
Textbook
▪ Convex Optimization, Stephen Boyd and Lieven Vandenberghe, Cambridge University Press

AI Robotics Lab. 2
Lecture Info.
공지사항 및 강의자료: 전북대 LMS
▪ [Link]

Grading
▪ Midterm Exam: 30%
▪ Final Exam: 30%
▪ Assignments: 20%
• 2 homeworks
• Late homeworks : 0점 처리
▪ Attendance: 10%
▪ Presentation: 10%
▪ Do not copy -> Direct F

AI Robotics Lab. 3
Lecture Goal
모든 것은 Problem Solving의 연속
본인만의 Problem Formulation → Optimization tool을 이용하여 해결
논문 Subsection 작성 완료

AI Robotics Lab. 4
Instructor
조형기 (HyungGi Jo)
▪ Assistant Professor, Division of Electronic Engineering
▪ AI Robotics Lab. ([Link]
▪ Office: 공과대학 7호관 417호
▪ Email hygijo@[Link]

AI Robotics Lab. 5
Convex Optimization?

AI Robotics Lab. 6
Convex Optimization Problem

두 이미지 간의 Correspondence matching 문제


• inlier/outlier가 존재함
• homography, fundamental matrix와 같은 geometric
transformation을 구해야 함

objective function 𝑆𝐼 : inlier set


𝜃: model parameters (homography, affine ...)
𝐴𝑖𝑗 : i번째 point와 j번째 point 간의 data association matrix
constraints
AI Robotics Lab. 7
Convex Optimization Problem
objective function 𝑆𝐼 : inlier set
𝐴𝑖𝑗 : i번째 point와 j번째 point 간의 data association matrix

constraints

𝑆𝐼 set의 cardinality를 최대화 시키는 𝜃(𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑚𝑎𝑡𝑟𝑖𝑥)와 𝑆𝐼 를 찾는다


→ consensus를 가장 크게 하는 set을 찾는다 ex) maximize the number of inliers
such that condition 1,2,3 ...

AI Robotics Lab. 8
Convex Optimization Problem
# Solving
Branch and Bound 적용시키면 LP (Linear Program) 문제로 바뀜

AI Robotics Lab. 9
Convex Optimization Problem

AI Robotics Lab. 10

You might also like