Emphysema detection in JPEG compressed HRCT lung images
Pramod Kumar Singh · 2006
We present an algorithm for the detection of emphysema pattern in the baseline JPEG Compressed HRCT lung imagesusingadecisiontreebasedcharacteristicconceptclassifier. The algorithm uses features extracted directly from DCT coefficients without its inverse transform. The extracted feature vector is used by concept rule sets derived using CRule, an extended characteristic decision tree algorithm, for the classification of 2x2 image blocks. The results of this algorithm in the detection of emphysematous patients are encouraging.