Active shape models—their training and application
Computer Vision and Image Understanding
Region Competition: Unifying Snakes, Region Growing, and Bayes/MDL for Multiband Image Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence
Computer Vision and Image Understanding
Layered Cooperation of Macro Agents and Micro Agents in Cooperative Active Contour Model
Agent Computing and Multi-Agent Systems
| Hi-index | 0.00 |
Active Contour Model, "Snake", is one of the most popular boundary detection methods. Its principle is an energy-minimizing spline for estimating the closest contour of a target object in an image gradually from an initial contour. However, this method has difficulty to determine an initial contour and parameters, and it cannot detect the target boundary precisely when the target image does not have clear edges or uniform feature. In this paper, we propose decentralized cooperative processing applied to Snake, which applied multiple Snakes to a single region, to improve its detection accuracy. The multiple Snakes run in coorperation with each other so as to increase the possibility of reaching the global optimum, and improve the estimation qualities. We verify the effectiveness of our proposal, in particular Multi-Snakes with different parameter sets, and Multi-Snakes applied to RGB-decomposed images, through the experiments using artificial images and real images. We then apply it to multi-spectral remote sensing, and show that our proposal detected the boundary with enough accuracy.