Thresholding of Biomedical Images
Venkatesan Rajinikanth, Nadaradjane Sri Madhava Raja, Nilanjan Dey · 2020
In the current age, the progression in science and engineering supports the human to have an improved livelihood ambiance with superior facilities. Even though therapeutic facilities are improved to provide the appropriate treatment process, the disease rate in humans is rising due to their lifestyle and change in the environmental conditions. Due to various satiations, the disease occurrence rates in humans are gradually increasing and a considerable number of disease preventive and treatment procedures also rising with the support of modern medical facilities. Modern technology also paved the way to detect the disease in premature stage and completely cure the disease with appropriate treatment procedures. The disease in internal body organs is acute compared to the disease in external organs. Further, noninvasive technique-based assessment of the disease is really essential to detect the disease with lesser efforts. The biomedical images recorded with a chosen modality play a vital role in noninvasive disease detection techniques, and based on the recorded image, the disease can be diagnosed by an experienced doctor or dedicated computer software. The recent literature confirms that the detection of disease using a chosen image processing technique considerably reduces the diagnostic burden. To ensure the accuracy during the disease detection based on the recorded image, it is necessary to include a considerable number of image pre-processing, post-processing, and decision making system. The overall result in the automated disease detection based on the biomedical image depends mainly on the initial result provided by the preprocessing system. If the outcome of the preprocessing scheme is poor, it degrades the final decision. Hence, it is necessary to consider a suitable preprocessing technique which will offer a better outcome. From the literature, it can be noted that the hybrid image examination techniques are widely employed in the literature to analyze medical images recorded with a class of modalities. The hybrid system can be formed by combining the preprocessing system with a carefully chosen post-processing system. In most of the cases, the preprocessing system employs a threshold operation implemented using a chosen OF and HA. In the medical image examination task, trilevel thresholding is widely executed to separate the test image into three sections, such as the background, normal image segment, and abnormal image segment. The abnormal section should be examined (ROI) by a post-processing section or by an experienced doctor. The evaluation result is then considered to take the decision regarding the abnormality and also this result plays a major role in the treatment planning process.