[Colloq] REMINDER (TODAY): Title: Towards holistic scene understanding for autonomous driving | Srikumar Ramalingam, Mitsubishi Electric Research Lab (MERL) | 2/22/16 2-3pm 366WVH

[Walker, Lashauna]

Title: Towards holistic scene understanding for autonomous driving

Speaker: Srikumar Ramalingam, Mitsubishi Electric Research Lab (MERL)

Date: 2/22/16

Time: 2:00-3:00pm

Location: 366 WVH



Abstract:
The success of an autonomous driving system (mobile robot, self-driving car) hinges on the accuracy and speed of inference algorithms that are used in understanding and recognizing the 3D world. A unifying theme of my research is the modeling of scene understanding problems via the lens of probabilistic graphical models and the development of new inference algorithms. In this talk, I will show novel algorithms for depth estimation and semantic segmentation from a single image. In particular, I will show a technique to solve the classical and ill-posed single-view 3D reconstruction problem using vanishing points, orthogonal structure, and a linear program that considers all plausible connectivity constraints between lines.
I will argue that by developing a principled and general framework for transforming problems with long-range interactions and complex dependencies (higher order multi-label functions) to simpler formulations (pairwise Boolean functions), we can use existing methods to solve these hard problems. In doing so, I will show transformations that are optimal with respect to some properties such as submodularity (discrete analogue of convexity). I will conclude by showing a few innovative and lightweight algorithms that achieve state-of-the-art results in the context of autonomous driving.



Short bio:

Srikumar Ramalingam is a senior principal research scientist at Mitsubishi Electric Research Lab (MERL) since 2008. He received a Marie Curie VisionTrain scholarship from European Union to pursue his doctoral studies in computer science and applied mathematics at INRIA Rhone Alpes (France). He received his PhD in 2007 under the guidance of Dr. Peter Sturm. His thesis on generic imaging models received INPG best thesis prize and AFRIF thesis prize (honorable mention) from the French Association for Pattern Recognition. He has published numerous papers in flagship conferences and journals (CVPR, ICCV, SIGGRAPH ASIA, ECCV, ICRA, IROS, RSS, TPAMI, and IJCV), won several awards, coauthored a book, co-organized international workshops, served as a guest editor for TPAMI, and given tutorials on topics such as graphical models, multi-view geometry, and non-classical cameras. His research interests are in computer vision, machine learning, robotics, and autonomous driving.






Thank You.

LaShauna Walker
Events and Administrative Specialist
College of Computer and Information Science
Northeastern University
617-373-2763
Facebook<https://www.facebook.com/ccisatnu?ref=hl> | Instagram<https://instagram.com/ccisatnu/> | LinkedIn<https://www.linkedin.com/groups/Northeastern-University-College-Computer-Information-1943637?gid=1943637&mostPopular=&trk=tyah&trkInfo=idx%3A1-1-1%2CtarId%3A1426606862845%2Ctas%3ANortheastern+University+College+of+Com> | Twitter<https://twitter.com/CCISatNU>