Student research opportunities

Image Segmentation and Scene Understanding

Project Code: CECS_817

This project is available at the following levels:
CS single semester, Honours, Summer Scholar, Masters

Keywords:

computer vision, machine learning

Supervisor:

Professor Stephen Gould

Outline:

Programming a computer to automatically interpret the content of an image is a long-standing challenge in artificial intelligence and computer vision. One way to interpret an image is via semantic segmentation, which is the task of automatically breaking an image into objects and regions and labelling each with a semantic class label (such as road, building, person, car, etc).

In this project you will work on cutting-edge computer vision and machine learning algorithms (including deep learning) to enhance the state-of-the-art in semantic segmentation.

Goals of this project

Develop algorithms and software for semantic segmentation. Work done in this project could lead to a scientific publication in a top quality conference or journal.

Requirements/Prerequisites

Strong programming skills in C++, Matlab and/or Python are required. A background in computer vision and machine learning is desirable.

Student Gain

Experience it cutting-edge research in machine learning and computer vision.

Background Literature

* http://users.cecs.anu.edu.au/~sgould/papers/cacm14-scene.pdf
* http://users.cecs.anu.edu.au/~sgould/papers/eccv14-spgraph.pdf
* http://users.cecs.anu.edu.au/~sgould/papers/cvpr12-multiSeg.pdf
* http://users.cecs.anu.edu.au/~sgould/papers/iccv09-sceneDecomposition.pdf

Contact: