Please use this identifier to cite or link to this item:
|標題:||THE EFFECT OF ROI NORMALIZATION FOR HAND RADIOGRAPHIC IMAGE SEGMENTATION||作者:||Lin, H.H.
|關鍵字:||Epiphyseal;Metaphyseal;2-means clustering;Bone age assessment;Regions of interest;Segmentation;bone-age assessment;regions;pyle||Project:||International Journal of Innovative Computing Information and Control||期刊/報告no：:||International Journal of Innovative Computing Information and Control, Volume 7, Issue 10, Page(s) 5669-5688.||摘要:||
The accuracy of epiphyseal/metaphyseal segmented results of a clustering scheme and the consistency of feature measurement depend on a reliable extraction of regions of interest (ROI) in an automatic bone age assessment system. ROI normalization is an important step for these purposes. The related modeling techniques are different according to the characteristics and applications of the processed images. Most of the epiphyseal metaphyseal ROI extraction schemes in the literature either only use hand normalization or do not use any normalization processing. Here, a novel ROI normalization scheme is proposed, which includes hand rotation to standard orientation as well as the individual finger rotation. An evaluation of the effect of the ROI normalization scheme in experiments derives from our distance approach adaptive 2-means clustering method and literature reviewed methods with regard to the final epiphyseal metaphyseal segmentation. Experimental results reveal that the proposed ROI normalization scheme provides a very well automated segmentation ability to separate accurately the epiphysis and metaphysis from the soft tissue of hand radiographs at the early stage of skeletal development. Furthermore, the experimental results show that the proposed approach provides a more stable performance for the segmentation of epiphyseal/metaphyseal regions.
|Appears in Collections:||資訊科學與工程學系所|
Show full item record
TAIR Related Article
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.