Histogram-based quantitative grading on ultrasonographic images of a fatty liver

Abstract

In this study, a new method called the Weighted Histogram (WH) was developed for enable the radiologist to interpret the level of the fatty liver objectively. Detection of fatty liver in the early stages is important for the control of the disease. Various imaging techniques such as MRI, CT and ultrasound are used for this purpose. However, ultrasound is the most widely used imaging technique in detecting fatty liver due to its non-invasive, affordable and easy accessibility. Figure A. Severe fatty liver, ultrasound imagePurpose: It is to develop a method that quantitatively determines the level of fatty liver using Ultrasound images to help the specialist by reducing the effects of subjective evaluation. With this method, the misdiagnosis caused by differences in the evaluation of fatty liver will be reduced.Theory and Methods: 256 level (8 bit) grayscale, 1024 x 1024 pixel digital liver ultrasound images has been used. A total of 4 regions, 20x20 pixels in size, are selected by the radiologist, one on the abdominal muscle for reference and three on the liver tissue for comparison. Histograms of the selected regions are calculated. The histograms obtained are recalculated by giving more weight to the white pixel so that new weighted histograms are obtained. To improve the ease of use of the proposed WH, a Graphical User Interface has been designed in the MATLAB environment. Results:The result showed that the outputs produced with using WH method are compatible with fatty liver grading in ultrasound images. In addition, the results more sensitive compared with the GRA method. The experimental results are presented and compared with GRA in Section 3. Conclusion:In the diagnosis of fatty liver by ultrasound, there is a subjective assessment based on the knowledge and visual perception of the radiologist. To increase objectivity and reduce operator dependency, in this study, a new quantitative assessment method was proposed to determine and grade the fatty liver from ultrasound images. In addition, the results obtained by using GRA were compared with the results obtained by the proposed method and the results are consistent. Also, WH was produced more sensitive results.