6-DoF Visual Localization for Autonomous Navigation

PhD Qualifying Examination


Title: "6-DoF Visual Localization for Autonomous Navigation"

by

Mr. Huaiyang HUANG


Abstract:

Recently, the problem of 6-Degree of Freedom (DoF) visual localization has been 
widely pursued in both computer vision and robotics communities. In the context 
of autonomous navigation for mobile robots, 6-DoF visual localization is a 
fundamental building block for other high-level applications, varying from 
global planning to decision making. However, the appearance and structure 
variances in the long-term real-world scenarios make this task challenging, 
requiring a further understanding of the prevailing methods and investigation 
of the open problems. Therefore, in this survey, we aim to provide a systematic 
overview of different aspects and paradigms and reveal the current progress and 
research gaps in 6-DoF visual localization.

This survey starts with a brief introduction to the task of 6-DoF visual 
localization and its abstract formulation. We first discuss the global and 
local data representations of visual data, which serve as a preliminary for 
major approaches. Through classifying different pipelines into four categories, 
namely retrieval-based and structure-based methods, scene coordinate 
regression, and direct pose regression, we broadly review prevailing methods 
and discuss their strengths and weaknesses. Then two tracks of method aiming to 
overcome the limitations are introduced, which either localize visual data over 
other modalities or embed the platform subtleties into the localization 
algorithm. To further analyze the performances of different methods and reveal 
the challenges of the long-term visual localization problem, we report and 
discuss the results of different methods on public benchmarks. Finally, we 
provide two promising directions and conclude this survey.


Date:			Tuesday, 9 June 2020

Time:                  	10:00am - 12:00noon

Zoom meeting:           https://hkust.zoom.us/j/91469775238

Committee Members:	Dr. Ming Liu (Supervisor)
 			Dr. Qifeng Chen (Chairperson)
 			Prof. Dit-Yan Yeung
 			Dr. Shaojie Shen (ECE)


**** ALL are Welcome ****