Multi – View Object Extraction with Fractional Boundaries
Abstract of Multi – View Object Extraction with Fractional Boundaries
There are two key contributions of our approach. First, we present an automatic method to identify a target object across different images for multi-view binary co-segmentation.
The extracted target object shares the same geometric representation in space with a distinctive color and texture model from the background.
Second, we present an algorithm to detect color ambiguous regions along the object boundary for matting refinement.
The local pixel band with the largest entropy is selected for matte refinement, subject to the multi-view consistent constraint.
Our results are high-quality alpha mattes consistent across all different viewpoints. We demonstrate the effectiveness of the proposed method using various examples.
Conclusion
In this Multi – View Object Extraction with Fractional Boundaries paper,
we introduced a framework to extract the soft boundaries of a target object from multi-view images.
We utilize coarse 3D reconstruction to define an initial volume bounding the foreground object.
Sequentially, we seek geometrically consistent regions having similar appearances across all input images.
The Fisher vector encoding adopted in the system allows us to model high-fidelity appearances in images.
To detect the optimal matte regions, we optimize the cumulative sum of KL divergences to smoothly take matte regions according to the contexts of object boundaries.
Our Laplacian matting equation considers geometrically consistent segmentations in enforcing the multi-view constraint for the final results.
