Active segmentation for robotics

The semantic robots of the immediate future are robots that will be able to find and recognize objects in any environment. They need the capability of segmenting objects in their visual field. In this paper, we propose a novel approach to segmentation based on the operation of fixation by an active observer. Our approach is different from current approaches: while existing works attempt to segment the whole scene at once into many areas, we segment only one image region, specifically the one containing the fixation point. Furthermore, our solution integrates monocular cues (color, texture) with binocular cues (stereo disparities and optical flow). Experiments with real imagery collected by our active robot and from the known databases [1] demonstrate the promise of the approach.

Authors

• Ajay K. Mishra
• Yiannis Aloimonos
• Cornelia Fermüller

Keywords

• Image segmentation
• Robot sensing systems
• Robot vision systems
• Robotics and automation
• Intelligent robots
• Layout
• Simultaneous localization and mapping
• Machine vision
• Navigation
• Orbital robotics
• Fixed Point
• Binocular
• Optical Flow
• Attention Mechanism
• Shortest Path
• Segmentation Algorithm
• Nodes In The Graph
• Line Segment
• Top-down Processes
• Bottom-up Processes
• Optimal Path
• Objects In The Scene
• Cartesian Space
• Segmentation Problem
• Flow Map
• Ground Truth Segmentation
• Edge Pixels
• Boundary Map
• Boundary Edges
• Polar Maps
• Internal Edges
• Polar Space
• Single Image
• Target Node
• Major Axis
• Object Of Interest
• Bilinear Interpolation
• Gradient Values
• Polar Coordinate System