2007 | OriginalPaper | Chapter
Improved incremental self-organizing map for the segmentation of ultrasound images
Authors : Zafer İşcan, Zumray Dokur, Tamer Ölmez
Published in: Mathematical Methods in Engineering
Publisher: Springer Netherlands
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
This paper presents an Improved Incremental Self-Organizing Map (I2SOM) network that utilizes automatic threshold (AT) value for the segmentation of ultrasound (US) images. I2SOM network has been compared with the well-known unsupervised Kohonen’s SOM network (KSOM) and a supervised Grow and Learn (GAL) network in terms of classification accuracy, learning time and number of nodes. For the feature extraction process, two-dimensional discrete cosine transform (2D-DCT) and 2D continuous wavelet transform (2D-CWT) were individually considered and were comparatively investigated to form the feature vectors of US breast and phantom images.
It is observed that the proposed automatic threshold scheme has significantly enhanced the robustness of I2SOM algorithm. Obtained results show that I2SOM can segment US images as good as Kohonen’s network.