310 likes | 450 Views
FOCUS : Clustering Crowdsourced Videos by Line-of-Sight. Puneet Jain , Justin Manweiler , Arup Acharya , and Kirk Beaty. Clustered by shared subject. c hallenges. CAN IMAGE PROCESSING SOLVE THIS PROBLEM?. Camera 1. Camera 2. LOGICAL similarity does not imply VISUAL similarity.
E N D
FOCUS: Clustering Crowdsourced Videos by Line-of-Sight Puneet Jain, Justin Manweiler, Arup Acharya, and Kirk Beaty
Camera 1 Camera 2 LOGICAL similarity does not implyVISUAL similarity Camera 3 Camera 4
Why not triangulate? Sensors are noisy, hard to distinguish subjects…
easy to identify hard to identify Don’t need to visually identify actual SUBJECT, can use background as PROXY Simplifying Insight 1
same basic structure persists Don’t need to directly match videos, can compare all to a predefined visual MODEL Simplifying Insight 2
Light-of-sight (triangulation) is almost enough, just not via sensing (alone) Simplifying Insight 3
Hadoop/HDFSFailover, elasticity Image processing Computer vision Clustered Videos Video Extraction Video Streams (Android, iOS, etc.) FOCUS Cloud Video Analytics Watching Live home: 2 away: 1 Change Angle Change Focus Users Select & Watch Organized Streams
Hadoop/HDFSFailover, elasticity Image processing Computer vision Clustered Videos Video Extraction FOCUS Cloud Video Analytics pre-defined reference “model” Watching Live home: 2 away: 1 Change Angle Change Focus Users Select & Watch Organized Streams
z z keypoint extraction multi-view reconstruction estimates camera POSE and content in field-of-view Multi-view Stereo Reconstruction Model construction technique based on Photo Tourism: Exploring image collections in 3D Snavely et al., SIGGRAPH 2006
z z keypoint extraction multi-view reconstruction z frame-by-frame video to model alignment ~ 18 seconds at 90th% ~ 1 second at 90th% z sensory inputs • Given a pre-defined 3D, align incoming video frames to the model • Also known as camera pose estimation
z z keypoint extraction multi-view reconstruction z integration of sensory inputs Gyroscope, provides “diff” from vision initial position Gyroscope 0 1 2 3 4 t - 1 t - 2 Filesize ≈ 1/Blur Sampled Frame
z z z keypoint extraction multi-view reconstruction pairwise model image analysis Field-of-view Using POSE + model POINT CLOUD, FOCUS geometrically identifies the set of model points in background of view
z z z keypoint extraction multi-view reconstruction pairwise model image analysis 3 Similarity between image 1 & 2 = 18 Similarity between image 1 & 3 = 13 2 1 Finding the similarity across videos as size of point cloud set intersection
Clustering “similar” videos Similarity Score • Application of Modularity Maximization • high modularity implies: • high correlation among the members of a cluster • minor correlation with the members of other clusters 1 1 2 2 3 3
Collegiate Football Stadium • Stadium 33K seats56K maximum attendance • Model: 190K points 412 images (2896 x 1944 resolution) • Android Appon Samsung Galaxy Nexus, S3 • 325 videos captured 15-30 seconds each
Line-of-Sight Accuracy GPS/Compass LOS estimation is <260 meters for the same percentage In >80% of the cases, Line-of-sight estimation is off by < 40 meters
FOCUS Performance 75% true positives Trigger GPS/Compass failover techniques
Natural Questions • What if 3D model is not available? • Online model generation from first few uploads • Stadiums look very different on a game day? • Rigid structures in the background persists • Where it won’t work? • Natural or dynamic environment are hard
Conclusion • Computer vision and image processing are often computation hungry, restricting real-time deployment • Mobile Sensing is a powerful metadata, can often reduce computation burden • Computer vision + Mobile Sensing + Geometry, along with right set of BigData tools, can enable many real-time applications • FOCUS, displays one such fusion, a ripe area for further research
Thank You http://cs.duke.edu/~puneet