Research Intern - Bayesian Methods in Geometric Computer Vision

Overview

Research Internships at Microsoft provide a dynamic environment for research careers with a network of world-class research labs led by globally-recognized scientists and engineers, who pursue innovation in a range of scientific and technical disciplines to help solve complex challenges in diverse fields, including computing, healthcare, economics, and the environment.

Traditional methods in geometric computer vision, which are still the go-to techniques for robot navigation, three-dimensional (3D) scene reconstruction, and the real-time generation of Augmented Reality/Virtual Reality (AR/VR) assets, are based on frequentist statistical methods, in particular maximum likelihood estimation. The goal of this Research Internship is to develop a basis for the application of Bayesian methods in geometric computer vision, starting with the fundamental problem of 2D and 3D point matching.

Candidates should have research-level knowledge of two of the three following topics (and, if required, the drive to learn the third), demonstrated through publications or explicit direction of their PhD work:

Geometric computer vision, including:

* The pinhole camera model, homographies, essential and fundamental matrix;

* three-dimensional point triangulation and iterative closest-point methods.

* Associated statistical methods such as RANSAC and maximum likelihood estimation;

* estimation of geometric and algebraic errors associated with geometric parameters of interest.

Bayesian statistical theory, including:

* Jeffreys' and uniform priors;

* exact and approximate marginalization techniques;

* nuisance parameters and outliers from a Bayesian perspective.

Rigorous methods in approximation theory, including:

* Polynomial randomized approximation algorithms;

* computational complexity classes;

* approximation methods for discrete problems.

Note that heuristic and learning-based methods are outside the scope of the work. Execution of the Research Internship plan will require efficient implementation of approximation algorithms in the context of one or more problems in geometric computer vision, and the benchmarking of the results against competing methods. Therefore, ability to produce efficient code in Python or C++ is a must.

The Research Internship has the explicit goal of producing a high-quality technical report, which, if the work is successful, will be submitted to an appropriate publication venue.

Responsibilities

Research Interns put inquiry and theory into practice. Alongside fellow doctoral candidates and some of the world's best researchers, Research Interns learn, collaborate, and network for life. Research Interns not only advance their own careers, but they also contribute to exciting research and development strides. During the 12-week internship, Research Interns are paired with mentors and expected to collaborate with other Research Interns and researchers, present findings, and contribute to the vibrant life of the community.

Research internships are available in all areas of research, and are offered year-round, though they typically begin in the summer.

Qualifications

Required Qualifications

* Currently enrolled in a PhD program in Mathematics, Statistics, Computer Science or a related STEM field.

* Completed graduate coursework in at least two of the three topics:

* Geometric computer vision;

* Probability and statistics, covering Bayesian theory;

* Approximation algorithms and computational complexity classes.

Other Requirements

* Research Interns are expected to be physically located in their manager's Microsoft worksite location for the duration of their internship.

* In addition to the qualifications below, you'll need to submit a minimum of two reference letters for this position as well as a cover letter and any relevant work or research samples. After you submit your application, a request for letters may be sent to your list of references on your behalf. Note that reference letters cannot be requested until after you have submitted your application, and furthermore, that they might not be automatically requested for all candidates.

You may wish to alert your letter writers in advance, so they will be…