TY - JOUR
T1 - Fused Multi-Characteristic Validity Index
T2 - An Application to Reconstructed Image Evaluation in Electrical Tomography
AU - Yue, Shihong
AU - Wu, Teresa
AU - Liu, Zhiqing
AU - Zhao, Xian
N1 - Funding Information:
This work was supported by the National Science Foundation of China under Grant No. 61174014, 60772080, 60532020 and the Tianjin Science Foundation of China under Grant No. 08JCYBJC13800.
PY - 2011/9
Y1 - 2011/9
N2 - The quality of reconstructed images is an important and direct criterion to quantitatively evaluate the effectiveness of reconstruction algorithms in electrical tomography (ET). Unfortunately, there lacks of effective and efficient approach to assessing the quality of ET images in literature and practices. Realizing the gap, we recently develop a novel index termed fused Multiple Characteristic Indices (fMCI) 1 aiming to measure the quality of clustering various data sets including imaging data (e.g., ET). In this paper, we propose a method based on fMCI to quantitatively evaluate the quality of reconstructed images. The method first applies the fast fuzzy c-means algorithm to cluster pixels in the reconstruct image. The fMCI is then applied to evaluate the clustering results and image quality. Simulated data derived from finite element method is used to demonstrate that the proposed method is capable to evaluate the quality of the reconstructed images and the results are consistent with visual perception. In addition, a number of experiments are conducted to demonstrate the applicability and effectiveness of the fMCI based method for image quality assessment.
AB - The quality of reconstructed images is an important and direct criterion to quantitatively evaluate the effectiveness of reconstruction algorithms in electrical tomography (ET). Unfortunately, there lacks of effective and efficient approach to assessing the quality of ET images in literature and practices. Realizing the gap, we recently develop a novel index termed fused Multiple Characteristic Indices (fMCI) 1 aiming to measure the quality of clustering various data sets including imaging data (e.g., ET). In this paper, we propose a method based on fMCI to quantitatively evaluate the quality of reconstructed images. The method first applies the fast fuzzy c-means algorithm to cluster pixels in the reconstruct image. The fMCI is then applied to evaluate the clustering results and image quality. Simulated data derived from finite element method is used to demonstrate that the proposed method is capable to evaluate the quality of the reconstructed images and the results are consistent with visual perception. In addition, a number of experiments are conducted to demonstrate the applicability and effectiveness of the fMCI based method for image quality assessment.
KW - Electrical tomography
KW - Validity index
KW - image evaluation
KW - image reconstruction
UR - http://www.scopus.com/inward/record.url?scp=84860858211&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84860858211&partnerID=8YFLogxK
U2 - 10.1080/18756891.2011.9727853
DO - 10.1080/18756891.2011.9727853
M3 - Article
AN - SCOPUS:84860858211
VL - 4
SP - 1052
EP - 1061
JO - International Journal of Computational Intelligence Systems
JF - International Journal of Computational Intelligence Systems
SN - 1875-6891
IS - 5
ER -