Enhancement of CT Images by Modified Object Based Contrast Stretching

The characteristic information of an organ lies in the texture of the image obtained by computed tomography imaging procedure. Therefore, enhancement of CT image aids a radiologist to better diagnose a disease which is a result of the enhanced/improved perceptibility and interpretability of the information present in the texture of the image. The present work proposes a method for texture enhancement of CT images by modifying Object-Based Multilevel Contrast Stretching method proposed by B. Xu et al. (IEEE Transaction on Consumer Electronics 3:1746-1754, 2010). In proposed method, a CT image is split into two images; (i) object approximation image and (ii) object error image. Object approximation image contains the overall structural information and object error image contains all the textural details of the CT image. The proposed method enhances the local contrast of object error image at an intra-object level, by greedy iterative stretch algorithm. Experimental results show that the proposed method enhances the textural details effectively, while maintaining the mean brightness of the image. Moreover, the quantitative results have verified that the proposed method outperforms the other methods.