China 
Africa 
Explore chapters and articles related to this topic
Application Development
Published in Scott E. Umbaugh, Digital Image Processing and Analysis, 2017
Bone cancer in dogs is a fatal disease that grows very fast on middle age to older dogs; both large breed as well as smaller breed [1]. Osteosarcoma (osteo = bone, sarcoma = cancer) is a common type of bone cancer that usually appear in the limbs which are called “appendicular osteosarcomas.” The nature of this kind of cancer is they grow from inside to outward and destroy the bone from the inside. A common symptom is lameness occurring one to three months after the first swelling is observed. The tumor that grows is not strong and breaks with slight injury and never heals. These fractures confirm the diagnosis of bone cancer and are commonly called “pathologic fractures.” Early detection of cancer is necessary before treatment is started. Analysis of radiography, MRI, and biopsies are the primary diagnostic tools for the detection of bone cancer. All of these techniques are not well suited for early diagnosis of disease both logistically and economically. A technique called IR thermography has been used as a tool for the preliminary detection of bone cancer in dogs. The primary advantages of this technique is that sedation is not necessary and changes in the thermographic pattern can be seen before they are actually confirmed using clinical signs and radiographic abnormalities [2,3]. The other advantages of using thermography are that the systems are portable, providing easy to use real-time imaging, and it is noncontact so it is hygienic and noninvasive in nature.
Level set evolution of biomedical MRI and CT scan images using optimized fuzzy region clustering
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2019
Kama Ramudu, Tummala Ranga Babu
Figure 3 shows an experimental result on 3T MRI brain scanning image; this image contains the three major tissues in the brain such as white matter (WM), grey matter (GM) and cerebrospinal fluid. These regions are very difficult to extract the details of the brain image for proper visualisation, size and diagnose a problem. Our proposed method is best suitable for accurate segmentation of tissues from the 3T MRI brain scanning. Figure 3(b)–(e) represents the clustered images with indexed value is 4. It means that the original image is divided into four regions. So each clustered image had the specific details of the object. Figure 3(f)–(g) showed that the selected regions are segmented using existing methods IVC2010 and 2013, respectively. Figure 3(h) showed that the selected portion is segmented using our proposed algorithm. The proposed algorithm segmented more regions compared with existing level set methods. Figure 4 represents an experimental result of photo micrographic finding of bone sarcoma which is to be segmented. Our algorithm segmented more regions over the existing techniques (see Figure 4(f)–(h)).Similarly, Figures 5 and 6 depicts the result of the Mammogram breast cancer Image and CT scan image of liver in which the tumour needs to be segmented, and Figure 7 shows the result of Alzheimer patient’s brain MRI scan. Compare to normal brain scan, Alzheimer patient’s brain MRI scan shows major brain shrinkage, and the relevant expansion of the ventricular system i.e. the black region at the centre of the image. IVC2010 and IVC2013 are the existing formulations which had drawback such as weak boundaries; selection limited to the brightest components only and eventually leading to a suboptimal segmentation. In contrast, the scheme of proposed OFR clustering-based segmentation was not only able to embark upon shortcomings of above two formulations, it in fact enhances objective indication variable hence showing the desired interest of selective level set segmentation. Figure 8 shows the final selected segmented regions using proposed level set approach using optimised fuzzy region (OFR) clustering of various real-time MRI and CT scans medical images. See Figure 9 for more example images to be simulated with existing and proposed models.