9-1-图像分割-概述.pdfVIP

图像识别与人工智能研究所,多谱信息处理国家重点实验室 图像分割 陶文兵 华中科技大学图像识别与人工智能研究所 多谱信息处理技术国家重点实验室 分割的目的和意义  图像分割是计算机视觉研究中的基础问题和难点之一  图像分割就是把图像分成各具特性的区域并提取出感兴趣目标 图像工程的三层模型  图像分割的难点和挑战性  对一般图像中的大量视觉模型进行建模的复杂性  图像理解本身的内在模糊性  当没有一个明确的任务来指导注意机制 2 image segmentation … Goal: Break up the image into meaningful or perceptually similar regions Types of segmentations Oversegmentation Undersegmentation Multiple Segmentations Segmentation as a result Rother et al. 2004 Segmentation for efficiency [Felzenszwalb and Huttenlocher 2004] [Shi and Malik 2001] [Hoiem et al. 2005, Mori 2005] Segments as primitives for recognition J. Tighe and S. Lazebnik, submitted to ECCV 2010 Major processes for segmentation Bottom-up: group tokens with similar features Top-down: group tokens that likely belong to the same object [Levin and Weiss 2006] Bottom-up segmentation • Group together similar-looking pixels for efficiency of further processing  “Bottom-up” process  Unsupervised “superpixels” X. Ren and J. Malik. Learning a classification model for segmentation. ICCV 2003. The goals of segmentation • Separate image into coherent “objects”  “Bottom-up” or “top-down” process?  Supervised or unsupervised? image human segmentation Berkeley segmentation database: /Research/Projects/CS/vision/grouping/segbenc