Static and Dynamic Image Segmentation and its applications in Vision-Based Control CS 294 -6 Advanced Topics in Computer Vision and Robotics
Vision-Based Landing of a UAV 2
Vision-based control of multiple vehicles Pursuit-Evasion Games Omni-based formation control 3
with a. ames Intensity-based Image Segmentation n n Apply GPCA to histogram or to eigenvector the of similarity matrix n=3 groups n n n Black Gray White n n=3 groups n n n Black Gray White 4
with a. ames Intensity-based Image Segmentation Time: 42 (sec) 70 (sec) 5
with a. ames Intensity-based Image Segmentation 36 (sec) 76 (sec) 3 (sec) 20 (sec) 35 (sec) 7
Texture-based Image Segmentation 8
Texture-based Image Segmentation 9
Texture-based Image Segmentation 10
Motion Segmentation: 2 views n n A static scene: multiple 2 D motion models n with s. soatto A dynamic scene: multiple 3 D motion models Given an image sequence, determine n n n Number of motion models (affine, Euclidean, etc. ) Motion model: affine (2 D) or Euclidean (3 D) Segmentation: model to which each pixel belongs 11
Segmentation of linear motions 12
Segmentation of linear motions n Transparent motion n Outline of a hand behind a lace curtain 13
Motion-based Image Segmentation 14
Experimental Results 15
Experimental Results 16
with o. shakernia Motion-based Image Segmentation n Omnidirectional Vision 17