YES - YEt another object segmentation: Exploiting camera movement

We addresvs the problem of object segmentation in image sequences where no a-priori knowledge of objects is assumed. We take advantage of robots' ability to move, gathering multiple images of the scene. Our approach starts by extracting edges, uses a polar domain representation and performs integration over time based on a simple dilation operation. The proposed system can be used for providing reliable initial segmentation of unknown objects in scenes of varying complexity, allowing for recognition, categorization or physical interaction with the objects. The experimental evaluation on both self-captured and a publicly available dataset shows the efficiency and stability of the proposed method.

Authors

• Lazaros Nalpantidis
• Mårten Björkman
• Danica Kragic

Keywords

• Image edge detection
• Image segmentation
• Object segmentation
• Correlation
• Optical imaging
• Robots
• Detectors
• Integration Time
• Scene Images
• Objects In The Scene
• Segmentation Problem
• Advantage Of The Ability
• Polar Representation
• Fixed Point
• Bounding Box
• Object Of Interest
• Segmentation Results
• Edge Detection
• 3D Volume
• Optical Flow
• Image Edge
• Information Propagation
• Consecutive Images
• Edge Information
• Polar Transformation
• Polarization Imaging
• Abstract Objects