MST Segmentation for Content-Based Medical Image Retrieval

Abstract

This paper describes an improved segmentation algorithm based on Minimum Spanning Tree (MST) for contentbased image retrieval system. MST segmentation is computationally efficient and captures both global and local image information, but it is prone to incur over-segmentation because of its neighbor system. To address this problem, an adaptive neighbor mode in the improved segmentation is defined by adding links between non-neighbor pixels of an image. The meaningful regions of an image are segmented automatically, and the region-based color features are exacted for the dominant segmented regions. The texture features are exacted using the Gabor filters, and are combined with the color features for retrieval The Experiments are performed using a medical database containing 370 images and the experimental results are shown and described finally.

Segmentation of Pre-processed Medical Images: An Approach Based on Range Filter

“…However, it is prone to cause the problem of over segmentation with some long and narrow redundant areas between two regions. Lu et al [16] describe an improved segmentation algorithm based on Minimum Spanning Tree (MST). To overcome the problem of over segmentation, an adaptive neighbor mode is defined by adding links between non-neighbor pixels of an image.…”

Medical Image Texture Segmentation Usingrange Filter

“…However, it is prone to cause the problem of over segmentation with some long and narrow redundant areas between two regions. In [2], authors describe an improved segmentation algorithm based on Minimum Spanning Tree (MST). To overcome the problem of over segmentation, an adaptive neighbour mode is defined by adding links between non-neighbour pixels of an image.…”