Hybrid image segmentation using watersheds and fast region merging

Abstract

A hybrid multidimensional image segmentation algorithm is proposed, which combines edge and region-based techniques through the morphological algorithm of watersheds. An edge-preserving statistical noise reduction approach is used as a preprocessing stage in order to compute an accurate estimate of the image gradient. Then, an initial partitioning of the image into primitive regions is produced by applying the watershed transform on the image gradient magnitude. This initial segmentation is the input to a computationally efficient hierarchical (bottom-up) region merging process that produces the final segmentation. The latter process uses the region adjacency graph (RAG) representation of the image regions. At each step, the most similar pair of regions is determined (minimum cost RAG edge), the regions are merged and the RAG is updated. Traditionally, the above is implemented by storing all RAG edges in a priority queue. We propose a significantly faster algorithm, which additionally maintains the so-called nearest neighbor graph, due to which the priority queue size and processing time are drastically reduced. The final segmentation provides, due to the RAG, one-pixel wide, closed, and accurately localized contours/surfaces. Experimental results obtained with two-dimensional/three-dimensional (2-D/3-D) magnetic resonance images are presented.

Learning object classes from structure

“…It can be thought of as measuring the volume of the symmetric difference of two classes. Its trace is proportional to the error as we define it, and is also related to the errors defined by Haris et al [15] and Malpica et al [17].…”

“…Unlike work aimed to segment images, see [15,10] for example our watersheds are computed directly from the intensity image, not from it derivative magnitude. Our grouping method is based on the work of Haris et al [15]. They define the cost of grouping a pair of regions as…”

“…Gauch [10] also segments by clustering watersheds of the derivative image, filtering the image in scale space to do so. Malpica et al [17] build on the work of Haris et al [15] by introducing multidimensional description for regions, rather than the scalar gray level. We also use a multi-dimensional description, but handle it differently from Malpica et al [17], as we will make clear below.…”

“…The arises in part because we seek to acquire objects classes in a semi-unsupervised environment, whereas watershed regions are usually grouped with a view to image segmentation. For example, Haris et al [15] use the watershed of the magnitude of the image derivative. They adopt a hierarchical agglomerative clustering approach which terminates when the cost of grouping any pair rises above a threshold (which varies from image to image).…”

Annotated Contraction Kernels for Interactive Image Segmentation

“…by merging adjacent regions that are rated similar by some appropriate cost measure (e.g. the difference of their average intensity) [1,2,3,4]. This bottom-up approach fits very well with the concept of irregular pyramids [5,6], and the main direction of this work is to show how the Active Paintbrush -an interactive segmentation tool developed for medical imaging [2] -and an automatic region merging [7,3,2] can be formulated based on the concepts of irregular pyramids and contraction kernels.…”