150 likes | 162 Views
Semantic Robot Vision Challenge. Ashutosh Kulkarni Shashank Senapaty Priyanka Singh March 17, 2008. Problem Statement.
E N D
Semantic Robot Vision Challenge Ashutosh Kulkarni Shashank Senapaty Priyanka Singh March 17, 2008
Problem Statement • Semantic Robot Vision Challenge (SRVC): Research competition designed to push the state of the art in image understanding and automatic acquisition of knowledge from large unstructured databases of images (such as those generally found on the web). • Use web to find image examples of objects • Identify these objects in a room environment Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Approach • Standard learning appearance models do not work • Unreliable training images CD “Hey Eugene” by Pink Martini • Solution: Rely on matching local features Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Basic Algorithm • Matched SIFT features between training and test images Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Bad (!) training images CD “Hey Eugene” by Pink Martini2.jpg Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Ranking (Google image search) Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Ranking (within group similarity) Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Problem with similarity based ranking Needed: DVD of Shrek 1 Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Elimination and Sub-classification • Weaker notion of a “good” training image – An image which matches with at least one other image from the class. • Remove images that do not match with any other image in the class Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
New Ranking results Good Images -- Class 1: Good Images --Class 2: Eliminated images: Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
But everything’s Harry Potter! False Positives: Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
RANSAC • Throw away outliers • Matching coefficient: Number of inliers / Total number of SIFT matches Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Results after RANSAC Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
References: • SRVC : http://www.semantic-robot-vision-challenge.org/ • Scott Helmer, David Meger, Per-Erik Forss´en, Tristram Southey, Sancho McCann, Pooyan Fazli, James J. Little, David G. Lowe. The UBC Semantic Robot Vision System. Webpage: http://www.cs.ubc.ca/~perfo/abstracts/hmfsfll07.html • Scott Helmer, David Meger, Per-Erik Forss´en, Sancho McCann, Tristram Southey,Matthew Baumann, Kevin Lai, Bruce Dow, James J. Little, David G. Lowe. Curious George: The UBC Semantic Robot Vision System. Webpage: http://www.cs.ubc.ca/~perfo/abstracts/hmfmsbldll07.html • Li-Jia Li, Gang Wang and Li Fei-Fei. OPTIMOL: automatic Object Picture collecTion via Incremental Model Learning. CVPR 2007. [PDF] • SIFT implementation: http://web.engr.oregonstate.edu/~hess/index.html • Web crawling resources: http://search.cpan.org/%7Egrousse/WWW-Google-Images-0.6.4/lib/WWW/Google/Images.pm Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh
Questions? Ashutosh Kulkarni • Shashank Senapaty • Priyanka Singh