An Approach Towards Effective Processing of Liver Tumour Data

Liver tumour has become one of the life-threatening
problems which need to be effectively diagnosed. Accurate
detection and segmentation of tumour region is crucial for
diagnosis. The major challenge persists with existing procedures
lack on false positive values. The objective of the work is to
preprocess the Liver CT images effectively in order to derive the
images capable of detecting the tumour infected area effectively.
The proposed novel work focused on creating an automated liver
CT image analysis through image processing technique. The
need for effective preprocessing is evaluated here. The study
considers MICCAI image dataset for evaluation to have a robust
preprocessing framework derived through Boundary extraction
algorithm (BEA). The purpose is to extract the region of interest
accurately through contour segmentation in order to
differentiate the liver structure from the background tissues. To
extract the required image intensities highlighted into the
extracted region, multi-scaling wavelet transform is applied. The
primary challenge of the system is to detect the abnormal pixels
present in the input CT images under test, accurately and
rapidly. The proposed novel Consistency aware natural looking
histogram equalization (CANHE) model is developed to enhance
the quality of the input image under test. The presented system
comprehensively evaluated to interpret the development of
performance measure while integrating the proposed CANHE
module and multi-scaling wavelet transform for feature
extraction.
Keywords— Medical imaging, Liver tumour, Image segmentation,
Computer vision, Image segmentation.