BAVM 2010 will take place at UC Berkeley on February 5th, 2010

The Bay Area Vision Meeting (BAVM) is an informal gathering (without a printed proceedings) of academic and industry researchers with interest in computer vision and related areas. The goal is to build community among vision researchers in the San Francisco Bay Area, however, visitors and travelers from afar are also encouraged to attend and present. New research, previews of work to be shown at upcoming vision conferences, reviews of not-well-publicized work, and descriptions of "work in progress" are all welcome.

Organizers:

• Trevor Darrell (UC Berkeley)
• Jitendra Malik (UC Berkeley)
• Mario Fritz (UC Berkeley)

Sponsors

Microsoft Research Sillicon Valley	Sony Electronics Inc.	Google Research
Intel	IBM Almaden Research Center	TYZX
	Like.com

Program

Invited Speakers:

• Pietro Perona (Caltech)
• Ruzena Bajcsy (UC Berkeley)
• Marc Levoy (Stanford)
• Luc Vincent (Google)
• Bruno Olshausen (UC Berkeley)

Agenda:

• 12:00-1:20pm Registration and Poster Setup
• 1:20-1:30pm Welcome (Organizers)
• 1:30-2:00pm Invited talk: Marc Levoy "Open-source cameras and the Stanford Frankencamera"
• 2:00-2:30pm Invited talk: Bruno Olshausen "What can biology teach us about vision?"
• 2:30-3:00pm Invited talk: Luc Vincent "Google Maps Street View: Overview & Computer Vision Challenges"
• 3:00-4:00pm Poster Session & Demos
• 4:00-4:30pm Invited talk: Ruzena Bajcsy "Tele-immersion, the cyber-infrastructure for studying body language"
• 4:30-5:00pm Invited talk: Pietro Perona "Visipedia"
• 5:00-6:00pm Poster Session & Demos

Posters:

1. Xiang Zhu: Non-Iterative Restoration for Weakly Blurred and Strongly Noisy Images
2. Pawan Mudigonda: Accurate MAP Inference for Region-Based Scene Understanding
3. David Chen: Inverted Index Compression for Scalable Image Matching
4. Jeff Mulligan: Robust Optical Eye Detection During Head Movement
5. Eugene Bart: Speeding Up Gibbs Sampling by Variable Grouping
6. Jimei Yang: Hierarchical Sparse Representation for Object Recognition
7. Alexandra Constantin, Ruzena Bajcsy, Sarah Nelson: Unsupervised Segmentation of Brain Tissue in Multivariate MRI
8. James Coughlan: Finding Small Bumps on the Ground: a Novel Formulation of Stereo Enforcing Smoothness of Elevation instead of Disparity
9. Ender Tekin: An Algorithm Enabling Blind Users to Find and Read Barcodes
10. Avideh Zakhor, George Chen and Matthew Carlberg: Classifying Urban Landscape in Aerial LiDAR Using 3D Shape Analysis
11. Daniel Leung: Proximate Sensing: Community Contributed Photographs as Volunteered Geographic Information
12. Daniela Ushizima and Jorge Cuadros: Image analysis of ocular fundus for retinopathy characterization
13. Yang Yi: Remote sensed image retrieval and classification using local descriptors
14. Hae Jong Seo: LARK: Locally Adaptive Regression Kernels as Visual Descriptors
15. Reza Moazzezi: Temporally distributed information gets optimally combined by change-based information processing
16. Silvio Savarese: Visual recognition in the three-dimensional world
17. Sanja Fidler: Evaluating multi-class learning strategies in a hierarchical framework for object
18. Adrien Bousseau: User Assisted Intrinsic Images
19. Mathieu Salzmann: Template-Free Monocular Reconstruction of Deformable Surfaces
20. Sergey Karayev: An Additive Latent Feature Model for Transparent Object Recognition
21. Brian Kulis: Learning to Hash with Binary Reconstructive Embeddings

Guideline for Poster Presentations

Please put up your poster before the talks start, so that it can be viewed during the full duration of the meeting. Your poster board will be tagged with the poster id as displayed in the program above. Both poster session will display all the poster. Please note that the maximum poster size supported by the supplied easels is 2' by 3'.

Demo Abstracts:

TYZX, Inc: TYZX G3 Embedded 3D Vision System:
TYZX will demonstrate their recently released DeepSea G3 Embedded Vision System (EVS). The G3 is a small, fast, rugged, smart 3D vision system that provides reliable, real-time range data for robots, security systems, and other applications. It is based on a new generation TYZX custom ASIC, the DeepSea 3, which integrates stereo processing with a ProjectionSpace processor, converting raw 3D point-cloud data into actionable data for applications in real time. The G3 EVS’ embedded Linux CPU is user-programmable for hosting applications such as obstacle detection and avoidance, person-following, and path-planning. The DeepSea 3 also features improvements in TYZX’s highly regarded stereo-correlation engine, extending both resolution and accuracy.

Like.com is the leading innovator in bringing computer vision and machine learning technology to the online fashion shopping space. This demo will focus on:
1. Photo realistic outfit creation. We are applying image based rendering techniques to help shoppers visualize their outfit from product images.
2. Unified search interface that combines browsing, text search, visual search and cross domain suggestions into a seemless experience.
3. User preference modeling and personalized search. We use advanced machine learning methods to model individual's style preferences to create a customized shopping experience without complex queries.
4. Visual style quiz based Covet.com product recommendation service.

Directions

The Bay Area Vision Meeting will take place in the Sutardja Dai Hall on the campus of UC Berkeley. For detailed directions see:

• Maps & Directions

Registration

The registration deadline for posters, demos, as well as attendance is January 22nd, 2010.

• Attendance registration form
• Poster registration form
• Demo registration form
• Parking reservation form (parking fee $15, reservation was available until January 29th, noon)

For comments and questions, please contact Mario Fritz at mfritz at eecs dot berkeley dot edu

Confirmed Attendees

Previous Meeting

BAVM 2009, Stanford