Pulling, Pushing, and Grouping for Image Segmentation
This paper presents a novel computational visual grouping method, termed pulling, pushing and grouping, or PPG for short. Visual grouping is formulated as a functional optimisation process. Our computational function has three terms, the first pulls similar visual cues together, the second pushes different visual cues apart, and the third groups spatially adjacent visual cues without regarding their visual properties. An efficient numerical algorithm based on the Hopfield neural model is developed for solving the optimisation process. Experimental results on various intensity, colour and texture images demonstrate the effectiveness of the new method.
KeywordsImage Segmentation Texture Image Computational Function Visual Property Visual Grouping
Unable to display preview. Download preview PDF.
- 1.Shi, J., Malik, J.: Normalized cuts and image segmentation. IEEE Trans. PAMI 22(8), 888–905 (2000)Google Scholar
- 2.Amir, A., Lindenbaum, M.: A generic grouping algorithm and its quantitative analysis. IEEE Trans. PAMI 20(2), 168–185 (1998)Google Scholar
- 3.Weiss, Y.: Segmentation using eigenvectors: a unifying view. In: ICCV (1999)Google Scholar
- 4.Randen, T., Husøy, J.: Filtering for Texture Classification: A Comparative Study. IEEE Trans. PAMI 21(4), 291–310 (1999)Google Scholar
- 5.Hopfield, J.: Neural networks and physical systems with emergent collective computational abilities. Proc. Natl. Acad. Sci. 2554–2558 (April 1982)Google Scholar
- 6.Hopfield, J.: Neurons with graded response have collective computational properties like those of two-state neurons. Proc. Natl. Acad. Sci. 3088–3092 (April 1984)Google Scholar
- 7.Hopfield, J., Tank, D.W.: Neural computation of decisions in optimization problems. Biol. Cyern., 141–152 (1985)Google Scholar