Title
Analysis of Ultrasound Kidney Images Using Content Descriptive Multiple Features for Disorder Identification and ANN Based Classification
Abstract
The objective of this work is to provide a set of most significant content descriptive feature parameters to identify and classify the kidney disorders with ultrasound scan. The ultrasound images are initially pre-processed to preserve the pixels of interest prior to feature extraction. In total 28 features are extracted, the analysis of features value shows that 13 features are highly significant in discrimination. This resultant feature vector is used to train the multilayer back propagation network. The network is tested with the unknown samples. The outcome of multi- layer back propagation network is verified with medical experts and this confirms classification efficiency of 90.47%, 86.66%, and 85.71% for the classes considered respectively. The study shows that feature extraction after pre-processing followed by ANN based classification significantly enhance objective diagnosis and provides the possibility of developing computer-aided diagnosis system.
Year
DOI
Venue
2007
10.1109/ICCTA.2007.31
ICCTA
Keywords
Field
DocType
pixel,biomedical imaging,image classification,backpropagation,testing,image analysis,neural nets,ultrasound,feature extraction,feature vector
Computer vision,Feature vector,Pattern recognition,Medical imaging,Computer science,Computer-aided diagnosis,Feature extraction,Kidney disorder,Pixel,Artificial intelligence,Backpropagation,Contextual image classification
Conference
ISBN
Citations 
PageRank 
0-7695-2770-1
6
0.86
References 
Authors
13
3
Name
Order
Citations
PageRank
K. Bommanna Raja1374.33
M. Madheswaran210215.57
K. Thyagarajah3354.19